The Anastomosis Angle Is a Key to Improved Long-Term Patency of Proximal Femoropopliteal Bypass
Article Outline
Background
Femoropopliteal bypass is a common vascular reconstructive procedure. A significant proportion of bypasses become ineffective within 1 year because of occlusion due to progression of intimal hyperplasia (IH).
Methods
The clinical part of the study involved an analysis of 43 patients with proximal femoropopliteal bypass, which became occluded no later than 1 year from the procedure, who were successfully treated with thrombolysis. Morphological changes of intima in the anastomosis (evaluated angiographically) and the angle of the distal end-to-side anastomosis were evaluated. In the second part of the study, blood flow in the distal end-to-side anastomosis was modeled experimentally (by particle image velocimetry) and numerically (by computational fluid dynamics). The results were correlated with the previously identified locations of IH.
Results
We proved that the locations of IH correlate with the locations of disturbed blood flow, increased wall shear stress, and stagnation points as documented by experimental visualization and angiographic findings. We also confirmed that anastomoses with more acute angles are less prone to IH and occlusion of the lumen.
Conclusion
We suggest that a better understanding of the hemodynamics and its influence on IH should lead to an optimized graft design by adopting a more acute angle of the anastomosis.
Introduction
Affecting nearly all known vascular reconstructive procedures, intimal hyperplasia (IH) constitutes an enormous clinical challenge. For the first time ever, in 1950, IH was documented in a canine model of arterial injury, described as “fibrous thickening of the intima” of microscopically evaluated specimens obtained 5-10 weeks after an operation of the femoral artery.1 Bypassing an affected section of the bloodstream is the most common method of vascular reconstruction. However, >25% of bypasses fail within the first year from surgery and >50% within the first 10 years. Stenosis at the site of the anastomosis due to IH is a principal cause of bypass failure within 1 year from the reconstructive procedure.2, 3
Many factors, with hemodynamics at the top of the list, contribute to the development of IH. Still, it remains unclear which type of blood flow shows the most harmful effect with respect to intimal proliferation. IH is associated with numerous flow parameters, e.g., turbulent flow, high or low or even oscillating wall shear stress (WSS), and high or low temporal or spatial WSS gradient.4, 5 Conversely, some authors have suggested that blood flow in an injured artery may have even a protective effect, acting purely as a mechanical factor that suppresses the endothelial response by washing away thrombogenic factors and cytokines.6
The objective of this study was to evaluate optimal geometrical parameters for an infrainguinal end-to-side anastomosis of a vascular bypass that would minimize the negative influence of hemodynamics on the vascular wall and therefore enhance its long-term patency. Moreover, we aimed to further our current understanding of the flow characteristics in such anastomoses.
Methods
In the clinical part of the study, 43 proximal (infrainguinal) femoropopliteal prosthetic bypasses, which became occluded within 6 months from reconstruction, were evaluated on a control angiography (AG) 1 day after successful thrombolysis. The anastomosis angle, changes in the vascular wall morphology suggestive of IH (narrowing of the lumen, irregular vessel wall), and percentage of residual flow in the host artery (proximal outlet segment [POS]) were quantified in the distal end-to-side anastomosis.
According to an analysis of the anastomosis angle and the frequency of intimal changes as evaluated on AG, there appeared to be two distinct groups of patients (group A angle ≤30°, n
=9; group B angle >30°, n=34). Changes suggestive of IH in the toe, the heel, and the floor of the anastomosis were evaluated; and the two groups were compared.
Afterward, experimental (flow visualization and particle image velocimetry [PIV]) and numerical (computational fluid dynamics [CFD]) models of the distal end-to-side anastomosis of a femoropopliteal bypass with complete obliteration of residual flow in the host artery were created and tested. The angle of the anastomosis was 25°, 45°, and 60° in each model type (Fig. 1). These values were chosen with regard to the distribution of angles found on AG (Fig. 2).
Fig. 1
Schematic drawing of the end-to-side anastomosis and regions of particular interest as recognized on AG. POS, proximal outlet segment; DOS, distal outlet segment.
Fig. 2
Flow velocity field in the 25° (top) and 60° (bottom) anastomosis models obtained from PIV. Detail of the anastomosis near the toe (right).
The fundamental principle of PIV is an evaluation of the velocity field from the shift of particles previously added to the working fluid (64% NaI solution in distilled water, ρ=1,730kg/m3, η=0.00254 Ns/m2, t=24°C). The fluid has the same refractive index as Plexiglas. The PIV system was supplied by Dantec Dynamics (Skovlunde, Denmark) and included a pair of cameras (Dantec HiSense 1,024×1,280 pixel charge-coupled device with a sampling frequency 4.5Hz for double-frame mode and 9Hz for single-frame mode), a pair of pulse lasers (Nd:YAG New Wave Gemini 15Hz/120 mJ with optics), and a PIV processor (Dantec FlowMap 1500 and 2×1 gigabit buffer). The motion was visualized by fluorescent particles (diameter 10μm) that emit light when exposed to an ND-YAG laser (λ=532nm).7
For the numerical simulation, CFD was used to model sine pulsatile flow in the anastomosis.8 With respect to the flow characteristics that were studied, the simulation was focused on the maximum of the pulse wave. In the maximum, the flow field is nearly identical to a stationary state with corresponding instant flow values.
An instance of anastomosis with 6mm internal diameter9 of both the graft and the host artery was modeled for a kinematic blood viscosity (ν=3.5 • 10−6 • m2 • s−1) and blood density (ρ=1,056kg • m−3).
The range of Reynolds number,
=–300 (negative sign indicates flow in a reverse direction) and ReMAX=1,000 in resting conditions.According to the similitude theory, transparent models with 20mm internal diameter of both the graft and the host artery were created. The models had rigid walls. As a working fluid, NaI water solution with added marking particles (diameter 10μm) was used. The visualization and PIV measurements were done in a model symmetry plane. For the numerical simulation (CFD), a three-dimensional computational hexahedral grid was generated. The length of the model distally from the anastomosis was 25 times the diameter of the model. For the simulation, a laminar mathematical model was chosen. Detailed descriptions of the experiment and the numerical methods can be found in the references.2, 7, 8, 10
Both the experiment and CFD were executed in four modes of stationary flow (Re=300, 500, 800, 1,000), and laminar flow was generated in the inlet into the anastomosis. The flow pattern was visualized and evaluated (vortices, stagnation points, etc.). In PIV the x and y axes' components of the mean velocity vector and their respective standard deviations were estimated:
To compare the distribution of WSS among the various models, WSS was normalized by the shear stress (τLAM), which was derived from a laminar profile with identical Reynolds number (Re).
is the mean volume velocity and ρ is the fluid density. The value of normalized WSS (WSSNORM) is calculated as
Results
In the first part of the study, AG in all 43 patients were evaluated. The anastomosis angle was below 30° in nine patients. Complete obliteration of the host artery was found in 37 patients. Changes in the intimal morphology were located in 34 patients in the toe, in 42 patients in the heel, and in 40 patients in the floor of the anastomosis.
According to the anastomosis angle (ranging between 20° and 60°) and the frequency of intimal changes on AG, two groups of patients were distinguished (group A angle ≤30°, n=9; group B angle >30°, n=34). Changes located in the toe of the anastomosis indicating neointimal proliferation were found in 11% in group A and in as much as 97% in group B (Table I), the only observation that proved significant (Fisher's exact test, p=0.036).
Table I. Tabular data grouped according to the angle of distal anastomosis as viewed on AG
| Anastomosis angle | Group size | Obliteration in POS | Intimal changes found in the following locations | ||
|---|---|---|---|---|---|
| Toe∗ | Heel | Floor | |||
| ≤30° | 9 | 8 (89%) | 1 (11%) | 9 (100%) | 7 (78%) |
| >30° | 34 | 29 (85%) | 33 (97%) | 33 (97%) | 33 (97%) |
In the experimental part of the study, we opted for a model of complete obliteration of the host artery (consistent with the finding of 86% of all anastomoses completely obliterated on AG). Moreover, no relationship between the presence of morphological changes detected on AG and the obliteration (partial versus complete) of the host artery was revealed (χ2, p>0.05) (Table II).
Table II. Tabular data grouped according to the degree of graft obliteration (complete or incomplete) as viewed on AG
| Obliteration in POS | Group size | Anastomosis angle >30° | Intimal changes found in the following locations | ||
|---|---|---|---|---|---|
| Toe | Heel | Floor | |||
| Complete | 37 | 29 (78%) | 30 (81%) | 37 (100%) | 34 (92%) |
| Incomplete | 6 | 5 (83%) | 4 (67%) | 5 (83%) | 6 (100%) |
Visualization of the flow field clearly delineated three zones of fluid stagnation (toe, heel, and floor) in the 45° and 60° models (Fig. 3, bottom), a finding that precisely corresponds to AG studies. Conversely, in the 25° model (Fig. 3, top), only two stagnation points (heel and floor) were identified. This again closely resembles the findings on AG.
Fig. 3
Distribution of the anastomosis angle: a histogram.
Evaluation of the velocity field by PIV revealed elevated WSS in the proximity of stagnation points in the 45° and 60° models, in contrast to the 25° model where no changes were observed. Moreover, more obtuse angles of anastomosis yielded higher values of flow velocity fluctuation in the toe, where the highest rate of changes on AG was observed (Fig. 4). Only slightly increased velocity fluctuation was found in anastomoses with more obtuse angles near the floor, and low values were observed in the heel, regardless of the anastomosis angle.
Fig. 4
Visualization of the 25° (top) and the 45° (bottom) anastomosis models in the symmetry plane. Detail of the anastomosis near the toe (right).
When comparing the distribution of WSS obtained from the numerical simulation of the 25° and 45° anastomosis models to the flow rate of fluid of arbitrarily set volume density (Re=300, 500, 800, 1,000), it became apparent that WSS was reduced in the case of more acute angles, particularly for higher values of Re. Similarly, WSSG was markedly decreased for the 25° angle compared to the 45° angle model, particularly in an area extending from the toe distally to x=2D (where D is the diameter of the model) (Fig. 5).
Fig. 5
Color map of fluctuation of flow velocity in the 25° (left) and the 60° (right) anastomosis models obtained from PIV.
Discussion
IH has been identified as a frequent cause of late failure of numerous revascularization procedures (bypass surgery, balloon angioplasty, atherectomy, etc.). This constitutes a serious problem of enormous scale. Our present understanding of its development remains somewhat incomplete, and there are no unified guidelines whatsoever as to how to prevent formation and slow down the progression of IH in clinical practice.
The accelerated growth of smooth muscle cells and the surrounding matrix is the principal cause of a long-term failure of end-to-side vascular graft anastomoses.11 Both clinical and experimental findings in previous studies demonstrated that the hemodynamics—flow pattern and WSS in particular—play a pivotal role in the development and localization of IH.12, 13, 14, 15, 16 However, the process alone is far more complicated. WSS magnitude and direction modulate cytoskeletal arrangement and expression of many growth factors. For example, nitric oxid synthase, transforming growth factor-beta, fibroblast growth factor 2, endothelin-1, platelet-derived growth factor B, angiotensin-converting enzyme, and vascular endothelial growth factor show a non-uniform change in expression according to WSS in the range of common WSS values (1-6dyn/cm2 in the venous system and 10-70dyn/cm2 in the arterial system).17, 18 In general, WSS below 5dyn/cm2 stimulates an atherogenic/vasoconstrictive profile, whereas values over 15dyn/cm2 induce endothelial quiescence; the endothelial response maintains mean arterial WSS magnitude at approximately 15-20dyn/cm2. High WSS values (70dyn/cm2 and more) have an adverse effect on endothelial function by causing mechanical injury.18 Several in vitro studies have investigated the influence of hemodynamics on the progression of anastomosis dysfunction. The importance of various characteristics such as the angle of the anastomosis,9, 19, 20, 21, 22 the symmetry and shape of the anastomosis,23, 24, 25 pulse shape, and inlet velocity profile have been researched.26 Some authors correlate anastomosis function with different flow characteristics like pressure and WSS magnitude,27 turbulence intensity, or disproportion between the diameter and compliance of the graft and the host artery.28, 29, 30 Globally decreased WSS is associated with a higher incidence of IH. This was clearly demonstrated by Binns et al.,31 who measured neointimal thickening in end-to-side polytetrafluorethylene (PTFE) graft anastomoses (diameter 4.6 and 8mm) in a canine model. They reported that oversized grafts induced lower WSS, resulting in a higher degree of IH, whereas undersized grafts led to an abnormally high WSS, with early thrombosis and graft failure due to mechanical endothelial injury with increased surface wettability. In another study, Kohler et al.32 demonstrated that increased blood flow and WSS resulted in a significant reduction of neointimal thickening in PTFE grafts in baboons. Similarly, Imparato et al.33 and Fry34 found that high-velocity flow and increased shear stress can effectively cause endothelial injury and, thus, sustain reactive subendothelial fibrocollagenous proliferation. Analysis of shear stress in an anastomosis led Faulkner et al.35 and Giordana et al.36 to believe that IH in distal anastomosis was a sequela of abnormal flow dynamics.
The question of the optimal angle of end-to-side anastomosis has been studied by numerical simulations (e.g., CFD)24, 25 and mechanical models (e.g., PIV),7, 22, 37 as well as in vivo studies.38, 39 The observations were notably unanimous: More acute angles imply decreased vortex and separation zones and reduced intimal thickening. In daily surgical practice, a distal end-to-side anastomosis with an angle below 20° is rarely constructed—the length of an anastomosis increases in sin(α)−1 proportion to the diameter of the graft, where α is the anastomosis angle. With a 20° angle the length of an anastomosis is already threefold the diameter of the graft. Moreover, it would be difficult to construct longer (more acute) anastomoses and embed them into their anatomic locations.
In our study, the evaluation of 43 AGs of femoropopliteal bypasses shortly after thrombolysis revealed a predilection for locations of intimal changes—predominantly in the toe, heel, and floor of the anastomoses. The significant difference in the frequency of changes in the toe of the anastomosis is indicative of the propensity of anastomoses with more obtuse angles to develop IH. This is not an unexpected result; however, it is this particular part of the study that supplies rare clinical evidence of the effect of anastomosis angle on the development of IH in the distal end-to-side anastomosis of a proximal femoropopliteal bypass in humans.
In the experimental part of the study, 25°, 45°, and 60° anastomosis models were used to visualize flow patterns by PIV (experimental model) and CFD (numerical model). In both models, the three locations (toe, heel, floor) in the anastomosis, previously identified on AG, were found to be similar to the locations that showed pathological hemodynamic parameters associated with IH—the presence of a stagnation point and WSSG.
Both models were visualized as shown in Fig. 3, Fig. 4, Fig. 5, Fig. 6, which provide a self-explanatory insight into the flow type. Immediately, it becomes clear that more acute anastomoses show laminar flow almost throughout the anastomosis, whereas more obtuse angles yield increasingly turbulent flow, especially in the aforementioned locations previously identified on AG with already described pathways leading ultimately to IH.
Fig. 6
Distribution of WSS and WSSG in the 25° (left) and the 45° (right) anastomosis models obtained from CFD for various flow rates (Re
=300, 500, 800, 1,000). All charts represent WSS or WSSG values as measured along a half-line beginning in the toe of the model, as indicated in the scheme. Nondimensional values on the x axis (x/D) represent the distance from the toe of the anastomosis (x) relative to the diameter of the tube (D).
The influence of WSS on the vessel wall depends to a considerable extent on the time factor and the area affected by WSS or WSSG. More importantly, WSS is related to the anastomosis angle, which can be optimized. Based on the results of this study and further literature data, we strongly believe that introducing the recommended angle of less than 45° (preferably ≤30°) of the distal femoropopliteal end-to-side anastomosis into everyday practice will prolong its life span.40
WSS can be experimentally assessed either indirectly by evaluating the flow velocity profile or directly using a special probe (constant temperature anemometry). In our experiment, we chose PIV, an indirect, noninvasive method, for the reason that it does not modify the flow by inserting a probe, like the direct methods do. The CFD model deliberately omits the fact that blood is a non-Newtonian fluid with complicated rheological and mechanical properties that are extremely complex and difficult to simulate. Moreover, we attempted to model nonstationary flow by stationary flow, and we had to choose several values of Reynolds number even though pulsatile flow cannot be represented by the Reynolds number alone. We used normalized WSS values to compare the models with each other, which was the objective. Due to the simplification of the model, it would be misleading to recalculate the values back to the instance. The model has a simplified geometry, neglecting the flexibility of a vessel wall and other factors like disproportion between the diameter and compliance of the graft and the host artery.
Conclusion
The first part of this study involved an analysis of AG findings of 43 successfully thrombolyzed proximal femoropopliteal bypasses that became occluded within 6 months from the reconstruction. Two groups with markedly different incidences of pathological changes in the vessel wall near the toe of the distal end-to-side anastomosis were identified. In the group with more acute angles (≤30°, n=9) changes in the toe occurred in 11% compared to 97% (p=0.036) in the group with more obtuse angles (>30°, n=34).
In the second part of the study, two models (physical [PIV] and numerical [CFD]) of a distal 25°, 45°, and 60° end-to-side anastomosis were created. Evaluation of both models according to flow patterns, stagnation points, WSS, WSSG, and velocity fluctuation near the vessel wall revealed a lower incidence of pathological changes in the anastomoses with more acute angles. These findings closely correspond with the changes found on AG.
This study supports the recommendation that angles below 45° (preferably ≤30°) are convenient for an end-to-side anastomosis and brings further evidence with an emphasis on the flow characteristics in an anastomosis obtained from experimental and computational models and, furthermore, correlated with human clinical data.
This research was supported by grant GA ČR 101/05/0675 (“Theoretical and Experimental Optimization of Vascular Reconstructions with Regard to Hemodynamics”).
APPENDIX.
Mean flow velocity in x and y axes (μu, μv) and their respective standard deviations (σu, σv) calculated from PIV measurement:
References
- . Arterial anastomosis in experimental arterial injury. Surg Gynecol Obstet. 1950;90:77–85
- Netrebska H, Schmirler M, Matecha J, et al. The stenosis shape influence on the flow pattern in distal end-to-side anastomosis. The 17th International Symposium on Transport Phenomena, Toyama, Japan, 2006.
- . The biology of saphenous vein graft occlusion: etiology and strategies for prevention. Curr Opin Cardiol. 1994;9:641–649
- . Infrainguinal graft surveillance: a biased overview. Semin Vasc Surg. 1993;6:143–149
- Hemodynamic parameters and early intimal thickening in branching blood vessels. Crit Rev Biomed Eng. 2001;29:1–64
- . Functional similarity of bypass arterial collateral vessel: studied on delayed occlusion in alternative arterial channels. Ann Surg. 1962;156:24–29
- Particle image velocimetry measurement in the model of vascular anastomosis. Prague Med Rep. 2007;108:75–86
- . Numerical study of the flow behind the end-to-side anastomosis. Seoul, Korea: World Congress on Medical Physics and Biomedical Engineering; 2006;pp 2698-2701
- Numerical model study of flow dynamics through an end-to-side anastomosis: choice of anastomosis angle and prosthesis diameter. Ann Vasc Surg. 2006;20:773–779
- Flow pattern and shear stress distribution in distal end-to-side anastomosis. Acta Bioeng Biomech. 2006;8:3–12
- Downstream anastomotic hyperplasia. A mechanism of failure in dacron arterial grafts. Ann Surg. 1983;197:479–483
- . Flow-mediated endothelial mechanotransduction. Physiol Rev. 1995;75:519–560
- . Intimal hyperplasia and hemodynamic factors in arterial bypass and arteriovenous grafts: a review. J Artif Organs. 2003;6:227–235
- Anastomotic intimal hyperplasia: mechanical injury or flow induced. J Vasc Surg. 1992;15:708–716
- . Intimal hyperplasia: an experimental study. Arch Surg. 1980;115:332–335
- Intimal hyperplasia and wall shear in arterial bypass graft distal anastomoses: an in vivo model study. J Biomech Eng. 2001;123:464–473
- . Shear stress magnitude and directionality modulate growth factor gene expression in preconditioned vascular endothelial cells. J Vasc Surg. 2003;37:182–190
- . Hemodynamic shear stress and its role in atherosclerosis. J A M A. 1999;282:2035–2042
- Effects of anastomotic angle on vascular tissue responses at end-to-side arterial grafts. J Vasc Surg. 2001;34:300–307
- . A numerical simulation of steady flow fields in a bypass tube. J Biomech. 2001;34:1407–1416
- . PIV measurements of proximal models with different anastomotic angles. Int Commun Heat Mass Transfer. 2000;27
- . Influence of angle on wall shear stress distribution for an end-to-side anastomosis. J Vasc Surg. 1994;19:1067–1073
- Influence of connection geometry and SVC-IVC flow rate ratio on flow structures within the total cavopulmonary connection: a numerical study. J Biomech Eng. 2002;124:364–377
- On the existence of an optimum end-to-side junctional geometry in peripheral bypass surgery—a computer generated study. Eur J Vasc Endovasc Surg. 2003;26:649–656
- . The influence of out-of-plane geometry on pulsatile flow within a distal end-to-side anastomosis. J Biomech. 2002;35:1225–1239
- Flow waveform effects on end-to-side anastomotic flow patterns. J Biomech. 1998;31:609–617
- . Experimental analysis of unsteady flows through a stenosis. J Biomech. 1998;31:11–19
- . Flow structures at the proximal side-to-end anastomosis: influence of geometry and flow division. J Biomech Eng. 1995;117:224–236
- . Improving vascular grafts: the importance of mechanical and haemodynamic properties. J Pathol. 2000;190:292–299
- . Particle hemodynamics analysis of Miller cuff arterial anastomosis. J Vasc Surg. 2003;38:1353–1362
- Optimal graft diameter: effect of wall shear stress on vascular healing. J Vasc Surg. 1989;10:326–337
- Increased blood flow inhibits neointimal hyperplasia in endothelialized vascular grafts. Circ Res. 1991;69:1557–1565
- Intimal and neointimal fibrous proliferation causing failure of arterial reconstructions. Surgery. 1972;72:1007–1017
- . Arterial intimal-medial permeability and coevolving structural responses to defined shear-stress exposures. Am J Physiol Heart Circ Physiol. 2002;283:H2341–H2355
- Effect of blood flow rate on subendothelial proliferation in venous autografts used as arterial substitutes. Circulation. 1975;52(2 Suppl.):I163–I172
- Local and global geometric influence on steady flow in distal anastomoses of peripheral bypass grafts 2005. J Biomech Eng. 2005;127:1087–1098
- Knapp Y, Bertrand E, Mouret BF. 2D-PIV measurements of the pulsative flow in a left heart simulator. Proceedings of Pacific Symposium on Flow Visualization and Image Processing (PSFVIP-4), Chamonix, France, June 3–5, 2003.
- An S-type bypass can improve the hemodynamics in the bypassed arteries and suppress intimal hyperplasia along the host artery floor. J Biomech. 2008;41:2498–2505
- . Axial flow fields in cuffed end-to-side anastomoses: effect of angle and disease progression. Eur J Vasc Endovasc Surg. 1999;18:240–244
- . Vascular Surgery. 6th ed.. Philadelphia: Elsevier Saunders; 2005;
PII: S0890-5096(09)00134-4
doi:10.1016/j.avsg.2009.06.008
© 2009 Annals of Vascular Surgery Inc. Published by Elsevier Inc All rights reserved.
