Journal of Biomechanics 37 (2004) 89–97

ARTICLE IN PRESS

*Correspondi

5803.

E-mail addre1Present Add

ing, Nanyang

Singapore 63979

0021-9290/$ - see

doi:10.1016/S002

Flow patterns in an endovascular stent-graft forabdominal aortic aneurysm repair

C.K. Chong1, T.V. How*

Department of Clinical Engineering, University of Liverpool, Duncan Building, Liverpool L69 3GA, UK

Accepted 27 May 2003

Abstract

Endovascular exclusion of the abdominal aortic aneurysm (AAA) has been carried out in selected patients during the past decade.

The deployment of a complex multicomponent endovascular device in an aneurysmal aorta may alter the local haemodynamics and

lead to thrombosis and intimal hyperplasia development. The aim of this in vitro study was to investigate the flow patterns using

flow visualisation and laser Doppler anemometry in a commercial bifurcated stent-graft. Two configurations of the stent-graft,

endo-stent and exo-stent, were investigated in an idealised planar AAA model. The flow structures in the main trunk in both

configurations of the stent-graft are three-dimensional with complex secondary structures. However, these flow structures were not

entirely caused by the stent-graft. The stent struts in the endo-stent configuration cause localised alteration in the flow pattern but

the overall flow structures were not significantly affected. Low velocity regions in the main trunk and flow separation in the stump

region and the curved segment of the iliac limbs were observed. These areas are associated with thrombosis in the clinical situation.

Improvements in the design of endovascular devices may remove these areas of unfavourable flow patterns and lead to better clinical

performance.

r 2003 Elsevier Ltd. All rights reserved.

Keywords: Flow visualisation; Laser Doppler anemometery; Flow velocity; Vascular prosthesis

1. Introduction

Abdominal aortic aneurysm (AAA) is a commonvascular disorder, which affects up to 5% of the malepopulation over 55 years of age in the West. Untilrecently, surgical replacement of the diseased aorticsegment with a prosthetic graft has been the standardform of treatment for AAA. Since the first clinical reportof endovascular exclusion of AAA using an intraluminalstent-graft device (Parodi et al., 1991), there has beenconsiderable progress and this type of operation is nowperformed routinely. A major advantage of the techni-que is that it is minimally invasive, recovery of thepatients is much quicker and obviates the need for cross-clamping of the aorta. The early results show that short-

ng author. Tel.: +44-151-706-5606; fax: +44-151-706-

ss: thienhow@liv.ac.uk (T.V. How).

ress: School of Mechanical and Production Engineer-

Technological University, 50 Nanyang Avenue,

8, Singapore.

front matter r 2003 Elsevier Ltd. All rights reserved.

1-9290(03)00236-7

term mortality and morbidity are comparable to thoseof the conventional treatment. The long-term clinicaleffectiveness and durability of the procedure and thecomplications related to the device itself are now beingreported in the literature. Stent-graft thrombosis andmicro-embolism are two complications associated withendovascular repair of AAA. In a recent report,Jacobowitz et al. (1999) showed that occlusion resultedin up to 11% of late explantation of stent-grafts whileseveral cases of fatal multi-organ failures have beenlinked to micro-embolism (Parodi, 1995). Although theexact mechanisms are not known, is has been suggestedthat the placement of stent alters the haemodynamicsand this coupled with wall movement may lead to thedispersion of late multiple emboli (Lindholt et al., 1998,Richter et al., 1999). The complex stent structures thatintrude into the blood flow may enhance biochemicalthrombotic cascade (Beythien et al., 1999) as well asdirectly affect the local haemodynamics (Peacock et al.,1995). Intimal hyperplasia is also associated with bloodvessel stenting although the consequences are moreserious in small diameter vessels. There is evidence that

ARTICLE IN PRESS

Fig. 2. The endovascular device, comprising of a main body and a

separate contralateral iliac limb, is shown with the original fabric cover

and the polyurethane cover. Note that the first two proximal stent

C.K. Chong, T.V. How / Journal of Biomechanics 37 (2004) 89–9790

the deployment of a stent causes flow disturbances whichmay contribute to neointimal thickening (Newman et al.,1996; Sukavaneshvar et al., 1998; Fabregues et al., 1998).Given the geometry of AAA devices and the non-uniformity and tortuosity of the aneurysm, blood flowdisturbances may be expected in endovascular stent-grafts. However, in order to gauge the role ofhaemodynamics in these complications, it is necessaryto define the flow characteristics of the stent-graft device.The aim of this in vitro study was to investigate the flowpatterns in a commercial stent-graft as a step towardsunderstanding the haemodynamic consequences of thedeployment of such a device.

segments are uncovered.

2. Materials and method

2.1. In vitro AAA model

A simplified planar model of an AAA was fabricatedfrom clear silicone rubber by rapid prototyping techni-que (Chong et al., 1999). The angulation of theinfrarenal aortic segment and tortuosity of the iliacand femoral arteries were not reproduced. Fig. 1 showsthe symmetrical AAA model with a straight segment ofthe aorta of diameter 25 mm, the aneurysm (max.diameter, 50 mm), the renal, iliac and femoral arteries(diameter 6, 12 and 10 mm, respectively).

The stent-graft used was a commercial modularbifurcated device (Vanguard Endovascular Aortic GraftSystem, Boston Scientific Ltd, St. Albans, UK). Duringintravascular deployment, the proximal part of thestraight contralateral iliac limb (diameter 12mm) islocated inside the tapered stump (minimum diameter9mm). To allow flow visualisation and laser Doppleranemometry (LDA) measurements the opaque fabriccover was replaced by thin-walled transparent polyur-ethane model (Fig. 2) of the same dimensions producedby a dip-coating process. The nickel–titanium (NiTi)stent structure of the device was retained.

In the commercial stent-graft device used here, the graftcomponent is sutured externally to the stent (endo-stentconfiguration). In other devices, however, the stent is fixedto the outside of the graft (exo-stent configuration). To

Fig. 1. Photograph of the simplified AAA model. It is shown with the

endovascular device in place.

investigate the effects of stent struts on the localhaemodynamics the two stent-graft configurations wereused. As in clinical practice, the device was deployed sothat the leading edge of the stent was distal to the lowestrenal ostium.

2.2. Flow circuit

The flow rig is shown in Fig. 3. The inlet section intothe model was a straight PVC pipe of diameter 25.4 mmand length 2m which incorporated an ultrasoniccannulating flow probe (24N, Transonic System Inc.,Ithaca, NY, USA), a calming chamber and pulsatilepump. The renal, common iliac and femoral arterieswere connected to the reservoir via four variable areaflowmeters provided with a needle valve (Perflow,Middlesex, UK). The actual mean flow rates weredetermined by measuring the volume of fluid collectedover a known time interval and the flowmeters wereused to provide an indication of the flow divisionthrough the outlets. Pulsatile flow was generated by aservo-controlled piston pump which provided theunsteady flow component and a gear pump (Micro-pump, Vancouver, WA, USA) which produced thesteady flow component. The simulated flow waveformrecorded in the suprarenal section by means of theultrasonic flow probe, shown in Fig. 4, was similar to thephysiological flow waveform measured by Moore andKu (1994).

Being constructed from a shape memory alloy, theNiTi stent become fully expanded only at bodytemperature. Therefore, it was necessary to maintainthe circulating fluid at 37�C. For flow visualisation, thefluid was a glycerol and sodium chloride solution towhich polyamide seeding particles (Rilsan, Elf AtochemLtd., UK) of diameter 75–150 mm were added. Thesolution had a dynamic viscosity of 3.62� 10�3 Pa s anda density of 1.13� 103 kg/m3 and. For LDA measure-ment, an aqueous solution of 46% (w/w) glycerol and14% (w/w) sodium chloride, with refractive index similar

ARTICLE IN PRESS

Time (s)

0.0 0.2 0.4 0.6 0.8 1.0

Flo

w (

l/min

)

0

1

2

3

4

a

c

d

e f

g

b

Fig. 4. Plot of the suprarenal flow waveform used in the experiments.

The waveform was recorded by means of a transit-time ultrasonic

flowmeter.

Fig. 5. Locations where flow visualisation and LDA measurement

were made. I suprarenal aorta, II infrarenal aorta, III proximal stent,

IV proximal trunk, V distal trunk, VI limb region, VII iliac vessels.

Fig. 3. Diagram of the flow circuit.

C.K. Chong, T.V. How / Journal of Biomechanics 37 (2004) 89–97 91

to that of the silicone rubber model was used. Thescattering particles were hollow glass particles of diameterB8mm (Spericel Hollow Spheres, Potters Industries Inc.,NJ, USA). The density of this solution was 1.16� 103 kg/m3 and its dynamic viscosity, 4.14� 10�3 Pa s.

Studies were carried out under simulated restingcondition at a heart rate of 60 beats/min and a meanaortic flow rate of 1.6 l/min with equal outflow of400 ml/min in each of the renal and iliac arteries. Thesevalues are in the physiological range and are comparableto those used in other in vitro studies of abdominalaortic flow (Moore and Ku, 1994; Pedersen et al., 1992).The corresponding mean Reynolds number (Re ¼rUD=m where D is the diameter of the vessel, U themean flow velocity and r and m are the density anddynamic viscosity of the fluid, respectively) and Wo-mersley number (a ¼ D=2O2pf r=m where f is thefrequency of the pulsatile flow) based on the aorticdiameter were 424 and 17.5, respectively for flowvisualisation studies, and those for LDA studies were379 and 16.6, respectively.

2.3. Flow visualisation

A 30 mW Helium–Neon laser provided with a linegenerator was used to illuminate the flow in the lateral(LR) and anterior–posterior (AP) planes. A videocamera (KP1-M, Hitachi Corp., Japan) was positionedat right angle to the laser sheet to record the flowpatterns traced out by the illuminated particles. Thecamera was operated at a frame rate of 25 frames/s andthe data were recorded onto S-VHS video tape via aframegrabber. The video data capture was synchronisedwith the piston pump so that each image frame can bereferenced to a known time period in the flow waveform.

2.4. LDA measurement

The LDA system used in the study was a two-component system (2D FibreFlow System, DantecElectronics Ltd., Bristol, UK). In the present study,only the axial velocity component will be presented.Velocity profile measurements were made in the LR andAP planes at the locations shown in Fig. 5, namely, thesuprarenal (I) and infrarenal (II) segments, the proximalstent (III), the proximal (IV) and distal (V) trunkregions, the left and right limb regions (VI) and thecurved regions of the iliac vessels (VII). However,complete profiles could not be obtained because of the

ARTICLE IN PRESSC.K. Chong, T.V. How / Journal of Biomechanics 37 (2004) 89–9792

obstruction of one or both laser beams by the stentstruts. Each velocity point represents an average of thedata from 128 consecutive cycles. The averaging processwas carried out using periodic trigger signals generatedby the pump system.

3. Results

3.1. Flow patterns

Flow structures at four phases of the flow cycle shownin Fig. 4, (a) the acceleration, (c) late deceleration, (d)end-systole, and (g) late diastole, at four locations in thestent are shown in Fig. 6. Since the video images wererecorded at 25 frames/s, each time point, (a)–(g),actually corresponds to a period of 40 ms. Supplemen-tary data in the form of movie images of the flowpatterns at the four locations can be viewed at theworldwide web address: http://www.jbiomech.com

3.1.1. Proximal stent region (Fig. 6, Row 1)

The flow patterns in the LR plane show that the fluidentering the uncovered portion of the device was

a c

1

LR

2

AP

3

LR

4

LR

Fig. 6. Flow patterns recorded in the proximal stent region, LR plane, (1); the

region, LR plane (4) recorded during acceleration (a), late deceleration (c),

images recorded in the LR plane and right to left in the AP plane. Movie im

http://www.jbiomech.com

undisturbed throughout the flow cycle. During thedeceleration phase of systole, a vortex appeared withinthe 3rd and 4th segment of the proximal stent. On thevideo recording, this vortex had a complex three-dimensional character. This vortex evolved into twocounter-rotating vortices at the end of systole, thenmerged again into a single structure which persistedthroughout diastole. In the proximal covered segmentthe flow was dominated by secondary structuresinvolving symmetrical counter-rotating vorticesthroughout the cardiac cycle.

In the AP plane (not shown), the laminar flow patternfrom the suprarenal region changed as the fluid enteredthe infrarenal region forming a central jet surrounded bythick separation layers at both the anterior and poster-ior walls. Particles near the anterior wall moved brieflyforward after mid-systolic acceleration and stoppedduring deceleration before reversing in end-systole andremaining stationary in mid-diastole. Particles near theboundary layer were seen to move into the mainstreamin diastole. Near the posterior wall, particles moved inthe reverse direction into the renal orifice in end-systolicdeceleration and late-systole. Particles near the bound-ary layer were swept forward by the incoming flow,

d g

trunk region, AP plane (2); the limb region, LR plane, (3); and the iliac

end-systole (d), and late diastole (g). Flow direction is left to right for

ages of the flow patterns can be viewed at the worldwide web address:

ARTICLE IN PRESSC.K. Chong, T.V. How / Journal of Biomechanics 37 (2004) 89–97 93

forming vortical structures in end-systole and continuedthroughout diastole. As the vortical structures began toclear after mid-systolic acceleration, wavy pathlineswere observed.

3.1.2. Trunk region (Fig. 6, Row 2)

Flow patterns recorded in the AP plane are shownhere. In this case the direction of flow is from right toleft due to the different orientation of the illuminationand the video camera. The trunk region is characterisedby complex flow patterns with evidence of instability insystolic acceleration phase, developing into a number ofvortical structures during systolic deceleration. Particlesin these structures either moved out-of-plane towardsthe anterior wall or circulated in the plane during thedeceleration phase and in end-systole. In late-diastole,these particles moved mainly out-of-plane towards theanterior wall. At the lateral walls, particles exhibited lowamplitude oscillatory motion, moving forward duringacceleration and diastole and backward in end-systoleand early diastole.

3.1.3. Limb regions (Fig. 6, Row 3)

The flow patterns in the LR plane are shown. In earlysystolic deceleration a region of flow separation wasobserved towards the right wall, immediately down-stream of the constriction caused by the presence of thecontralateral limb in the stump. This developed into atransient and well defined and vortex in end-systole,pushing the streamlines outwards and resulted in theformation of another smaller separation region down-stream, near the left lateral wall. This pattern was alsoseen in the AP plane (not shown). In the right limb,however, the flow was relatively undisturbed.

3.1.4. Iliac regions (Fig. 6 Row 4)

The flow patterns shown were recorded in the LRplane. During late-systolic acceleration and peak-systolein the left iliac artery the flow was skewed towards theinner curved wall with flow separation occurring at theouter wall. During the deceleration phase the flowseparation region grew to occupy more than half of thediameter of the stent-graft in end-systole. Within thisregion, particles exhibited a spiral motion, arisinganteriorly from the inner wall and descending poster-iorly into the main stream.

3.2. Velocity profiles

3.2.1. Endo-stent configuration

Fig. 7 shows the velocity profiles obtained using LDAin the LR (left column) and AP (right column) planes atvarious positions along the device. At each position,velocity profiles at 7 points in the flow cycle, (a)–(g), asshown on Fig. 4 are presented. In the suprarenal region

(not shown), the velocity profiles were generallysymmetrical about the centreline in both planes.

3.2.2. Infrarenal aortic neck region (II)

The profiles in the LR plane were skewed slightlytowards the left with the peak velocity, 19.5 cm/s,located at about 1 mm from the central flow axis. Inthe AP plane, the velocity profiles after peak systoleassume a W shape. As the flow cycle progressed, theregion of flow reversal extended further from the wallsand in late-diastole, negative velocities were found at upto 7mm from the walls.

3.2.3. Proximal stent region (III)

In the LR plane, the velocity profiles appeared tobecome more uniform throughout the cardiac cycle.Because of obstruction of the laser beams, measure-ments near the walls could not be obtained. In the APplane, flow reversal was apparent at the walls in late-systolic deceleration but significant parts of the lumenremained obscured by the stent struts.

3.2.4. Proximal trunk region (IV)

In the LR plane the velocity profile was fairly uniform(18–19.5 cm/s) at peak-systole. In late-diastole, flowreversal and stagnation occurred near the lateral wallsand in an area stretching 6mm on either side of the axis.In the AP plane, central streaming of the flow continuedfrom the proximal stent region with an indication of W-shaped profiles after peak systole. However, velocitymeasurements could not be obtained from a regionabout 9mm from the posterior wall.

3.2.5. Distal trunk region (V)

Further downstream, the M-shaped velocity profilesbecame more prominent, at least near the right wallwhere data were available. In late systolic decelerationand late-diastole flow stagnation, retrograde, andantegrade flows occurred over the central region andnear the right wall. Insufficient data were available inthe AP plane to define the velocity profiles.

3.2.6. Right and left limb regions (VI)

The flow in both limbs were triphasic, with a largeretrograde component in end-systole. In the LR plane,the velocity profiles appeared to be symmetrical aboutthe axis, being relatively flat in acceleration and peaksystole changing into M- and W-shaped profiles as thecycle progressed. Peak velocities were higher in the leftlimb.

Data could only be obtained from the anterior half ofthe AP plane but the available data suggest that thevelocity profiles are similar to those in the LR plane.Again the peak velocity was higher in the left limb.

ARTICLE IN PRESS

Fig. 7. Axial velocity profiles in the central LR (left column) and AP (right column) planes; endo-stent configuration. The profiles were obtained at

the locations as shown in Fig. 5 and are shown at different times in the flow cycle: (a) yellow, (b) red, (c) purple, (d) blue, (e) black, (f) turquoise, (g)

green. The horizontal axis denotes the radial position in mm and 0 corresponds to the centreline.

C.K. Chong, T.V. How / Journal of Biomechanics 37 (2004) 89–9794

ARTICLE IN PRESS

Fig. 8. Axial velocity profiles in the central LR (left column) and AP (right column) planes; exo-stent configuration. The profiles were obtained at the

locations as shown in Fig. 5 and are shown at different times in the flow cycle: (a) yellow, (b) red, (c) purple, (d) blue, (e) black, (f) turquoise, (g)

green. The horizontal axis denotes the radial position in mm and 0 corresponds to the centreline.

C.K. Chong, T.V. How / Journal of Biomechanics 37 (2004) 89–97 95

ARTICLE IN PRESSC.K. Chong, T.V. How / Journal of Biomechanics 37 (2004) 89–9796

3.2.7. Right and left iliac regions (VII)

The velocity profiles (Fig. 8) in the two iliac vesselswere similar but there were clear differences between thetwo planes. The flow was highly asymmetric in the LRplane, skewing markedly towards the right or inner wallwhile at the left, significant retrograde flow occurred insystolic deceleration and diastole, resulting in extensiveflow separation covering over half the diameter of theiliacs. The velocity profiles were highly irregular insystolic deceleration and early diastole. In the AP plane,the flow was nearly more symmetrical about the vesselaxis. In late-systolic deceleration and diastole the profilehad a pronounced M-shape.

3.3. Exo-stent configuration

The velocity profiles were generally similar to those ofthe endo-stent case, although they appeared to besmoother. As in the endo-stent case, the profiles in theaortic and proximal stent regions were more uniform inthe LT than in the AP planes. The M-shaped profiles inthe trunk regions in the LR planes were better defined.In the AP plane however, flow in the posterior half ofthe main trunk could not be measured because of lack ofoptical access.

More velocity points were obtained in the limbregions, particularly in the AP plane. The velocityprofiles appeared to be well organised in the right limbin the LR plane. In the left limb however, there weresignificant differences in the shape of the profilescompared with the endo-stent case during the secondhalf of the pulsatile flow cycle.

4. Discussion

The deployment of a complex multicomponentendovascular device in the vascular system is likely toalter the local haemodynamics and adversely affect thelong-term performance of the device. Although previousmodel studies have been performed on the effect ofcoronary (Peacock et al., 1995) and iliac (Fabregueset al., 1998) stents on flow structures, there does notappear to have been similar studies on AAA devices.

In this first study, an idealised planar AAA modelwith a long inlet length was used to ensure fullydeveloped and stable flow. The flow visualisation studiesreveal that even in this ideal case the flow patternswithin the stent-graft were highly three dimensional withcomplex secondary structures. These flow features werefirst attributed to the presence of the stent-graft and thegeometry of the specific device under study. However,similar flow structures have been observed in this regionof the normal human circulation system (Moore et al.,1992; Moore and Ku, 1994; Pedersen et al., 1992, 1994).These flow disturbances in the infrarenal abdominal

aortic segment are probably related to the large outflowthrough the renal arteries, resulting in a reduction invelocity and an adverse pressure gradient. The velocityprofiles in the main trunk of the stent-graft show thisgeneral reduction of velocity as the flow progressed fromthe proximal stent towards the distal trunk. Similar flowstructures have also been reported by Rieu and Pelissier(1991) in their study of flow in bifurcated aortofemoralvascular prosthesis. Low velocities and low amplitudevelocity oscillations were observed near the walls of thegraft. These may be due to small troughs or furrows in thegraft wall caused by the attachment of the flexible andcompliant graft to the rigid sent structure. These regionsof low velocities may be prone to thrombus formation.

In the limb regions, flow separation extending overhalf the vessel diameter was observed in the left limbdistal to the stump. These flow disturbances can beattributed to the modular design of bifurcated stent-graft, which results in a constriction at the stump (anumber of other modular devices are currently in use).In the right limb where no constriction was present, nosuch flow separation was seen at any time during thecardiac cycle. Fixation of the left limb relies on theradial force that it exerts on the undersized stump. Inthis specific design of modular stent-graft, the constric-tion resulted in an estimated 45% reduction in cross-sectional area. Distal to this constriction, or stenosis, theflow separation region was skewed towards the rightside as a result of the asymmetry in the geometry. Theregion of flow separation is known to be associated withaggregation of platelets and accumulation of activationfactors (Bluestein et al., 1999). In this study the particlesin the separation region appeared to have long residencetime, and this may contribute to thrombus formation(Richter et al., 1999; Sukavaneshvar et al., 1998). Inaddition, the thrombus formation may be enhanced bythe presence of the metallic stent in contact with blood(Gutensohn et al., 1997; Sukavaneshvar et al., 2000). Itmay be possible also that shear-induced activation ofplatelet could occur in the stenosis (Sakariassen et al.,1998). These may explain the clinical cases where stent-grafts of this specific design failed due to thrombosis inthis region (Stelter et al., 1997; Becquemin et al., 1999;Umscheid and Stelter, 1999).

In both the right and left iliac vessels, the flow wascharacterised by skewing of the flow towards the innerwall and extensive flow separation towards the outerwall, which covered more than half the vessel diameter.These are classic flow features found in curved vessels(Dean flow). The disturbed flow patterns in the iliacregion may partly explain the clinical cases of stent-grafts of the specific design used in this study (Stelteret al., 1997; Becquemin et al., 1999; Umscheidand Stelter, 1999) and others, e.g. the EVT device(Jacobowitz et al., 1999) that reported to have failedbecause of thrombosis in the limb region.

ARTICLE IN PRESSC.K. Chong, T.V. How / Journal of Biomechanics 37 (2004) 89–97 97

One objective of this study was to investigate theeffect of the stent struts on the flow pattern. This wasachieved by comparing the flow pattern and velocityprofiles between the exo- and endo-stent configurationsin the AAA model. The flow visualisation resultsindicate that there was no significant difference in theoverall flow pattern between the two cases. However,the velocity profiles in both the LR and AP planes in thecentral portion of the flow axis show that the exo-stentcase appeared to be smoother and more stable thanthose in the endo-stent case. These observations suggestthat the effect of the stent on overall haemodynamicswithin the main trunk is small and this is probably dueto the stent strut-to-main trunk diameter ratio beinglow. Flow patterns in the immediate vicinity of the stentstruts were also observed. Particles were trapped in theregions where the stent segments are joined, particularlyat the proximal neck where the stent density is high andthe stent diameter largest (0.4 mm). The velocity in thisregion was low, and the introduction of stent furthercreated low velocity zones. In the endo-stent case wherestent struts intruded into the flow the particle pathlinesaround the stent struts could be seen on the videorecordings. This was not observed in the exo-stent case.

The influence of this specific stent-graft geometry onthe flow pattern can also be appreciated from changes inthe overall velocity profiles. The nearly parabolicvelocity profiles became much flatter in the LR planeas the flow entered the stent-graft while centralstreaming occurred in the AP plane. Downstream inthe limb and iliac vessels, plug flow was observed in bothplanes. The change in the velocity profiles was enhancedby the tapering of the stent-graft in the AP plane fromthe proximal trunk towards the stent-graft bifurcationand an expansion in the diameter just before thebifurcation in the LR plane. This resulted in a lowerpressure gradient as evidenced by the gradual decreasein velocity along the main stent-graft component,resulting in the central flattening of the velocity profiles.These flow features were similar to those presented byRieu and Pelissier (1991) in the LR plane. A fullcomparison, however, may not be possible since they didnot study the flow in the AP plane.

References

Becquemin, J.P., Lapie, V., Favre, J.P., Rousseau, H., 1999. Mid-term

results of a second generation bifurcated endovascular graft for

abdominal aortic aneurysm repair: the French Vanguard trial.

Journal of Vascular Surgery 30, 209–218.

Beythien, C., Gutensohn, K., Bau, J., Hamm, C.W., Kuehnl, P.,

Meinertz, T., Terres, W., 1999. Influence of stent length and

heparin coating on platelet activation: a flow cytometric analysis in

a pulsed floating model. Thrombosis Research 94, 79–86.

Bluestein, D., Gutierrez, C., Londono, M., Schoephoerster, R.T.,

1999. Vortex shedding in steady flow through a model of an arterial

stenosis and its relevance to mural platelet deposition. Annals of

Biomedical Engineering 27, 763–773.

Chong, C.K., Rowe, C.S., Sivanesan, S., Rattray, A., Black, R.A.,

Shortland, A.P., How, T.V., 1999. Computer aided design and

fabrication of models for in vitro studies of vascular fluid dynamics.

Proceedings of Institution of Mechanical Engineers Part H 213, 1–4.

Fabregues, S., Baijens, K., Rieu, R., Bergeron, P., 1998. Hemodynamics

of endovascular prostheses. Journal of Biomechanics 31, 45–54.

Gutensohn, K., Beythien, C., Bau, J., Meinertz, T., Kuehnl, P., 1997. Flow

cytometric analysis of coronary stent-induced alterations of platelet

antigens in an in vitro model. Thrombosis Research 86, 49–56.

Jacobowitz, G.R., Lee, A.M., Riles, T.S., 1999. Immediate and late

explantation of endovascular aortic grafts: the endovascular

technologies experience. Journal of Vascular Surgery 29, 309–316.

Lindholt, J.S., Sandermann, J., Bruun Petersen, J., Nielsen, J.O.,

Fasting, H., 1998. Fatal late multiple emboli after endovascular

treatment of abdominal aortic aneurysm. Case report. Interna-

tional Angiology 17, 241–243.

Moore Jr., J.E., Ku, D.N., 1994. Pulsatile velocity measurements in a

model of the human abdominal aorta under resting conditions.

ASME Journal of Biomechanical Engineering 116, 337–346.

Moore, J.E., Ku, D.N., Zarins, C.K., Glagov, S., 1992. Pulsatile flow

visualization in the abdominal aorta under differing physiologic

conditions: implications for increased susceptibility to atherosclero-

sis. ASME Journal of Biomechanical Engineering 114, 391–397.

Newman, V.S., Berry, J.L., Routh, W.D., Ferrario, C.M., Dean, R.H.,

1996. Effects of vascular stent surface area and hemodynamics on

intimal thickening. Vascular Interventional Radiology 7, 387–393.

Parodi, J.C., 1995. Endovascular repair of abdominal aortic aneurysms

and other arterial lesions. Journal of Vascular Surgery 21, 549–555.

Parodi, J.C., Palmaz, J.C., Barone, H.D., 1991. Transfemoral

intraluminal graft implantation for abdominal aortic aneurysms.

Annals of Vascular Surgery 5, 491–499.

Peacock, J., Hankins, S., Jones, T., Lutz, R., 1995. Flow instabilities

induced by coronary artery stents: assessment with an in vitro pulse

duplicator. Journal of Biomechanics 28, 17–26.

Pedersen, E.M., Yoganathan, A.P., Lefebvre, X.P., 1992. Pulsatile flow

visualization in a model of human abdominal aorta and abdominal

aortic bifurcation. Journal of Biomechanics 25, 935–945.

Pedersen, E.M., Sung, H.W., Yoganathan, A.P., 1994. Influence of

abdominal aortic curvature and resting versus exercise conditions

on velocity fields in the normal abdominal aortic bifurcation.

ASME Journal of Biomechanical Engineering 116, 347–354.

Richter, G.M., Palmaz, J.C., Noeldge, G., Tio, F., 1999. Relationship

between blood flow, thrombus, and neointima in stents. Journal of

Vascular Interventional Radiology 10, 598–604.

Rieu, R., Pelissier, R., 1991. In vitro study of a physiological type flow

in a bifurcated vascular prosthesis. Journal of Biomechanics 24,

923–933.

Sakariassen, K.S., Holme, P.A., Orvim, U., Barstad, R.M., Solum,

N.O., Brosstad, F.R., 1998. Shear-induced platelet activation and

platelet microparticle formation in native human blood. Throm-

bosis Research 92 (Suppl. 2), S33–S41.

Stelter, W., Umscheid, T., Ziegler, P., 1997. Three-year experience with

modular stent-graft devices for endovascular AAA treatment.

Journal of Endovascular Surgery 4, 362–369.

Sukavaneshvar, S., Solen, K.A., Mohammad, S.F., 1998. Device

induced thromboembolism in a bovine in vitro coronary stent

model. ASAIO Journal 44, M393–396.

Sukavaneshvar, S., Rosa, G.M., Solen, K.A., 2000. Enhancement of

stent-induced thromboembolism by residual stenoses: contribution

of hemodynamics. Annals of Biomedical Engineering 28, 182–193.

Umscheid, T., Stelter, W.J., 1999. Time-related alterations in shape,

position, and structure of self-expanding, modular aortic stent-

grafts: A 4-year single-center follow-up. Journal of Endovascular

Surgery 6, 17–32.