Finite element analysis of stent deployment: understanding stent fracture in percutaneous pulmonary valve implantation.

Abstract

To analyze factors responsible for stent fracture in percutaneous pulmonary valve implantation (PPVI) by finite element method.
PPVI is an interventional catheter-based technique for treating significant pulmonary valve disease. Stent fracture is a recognized complication.
Three different stent models were created: (1) platinum-10% iridium alloy stent - resembles the first-generation PPVI device; (2) same geometry, but with the addition of gold over the strut intersections - models the current stent; (3) same design as 1, but made of thinner wire. For Model 3, a stent-in-stent solution was applied. Numerical analyses of the deployment of these devices were performed to understand the stress distribution and hence stent fracture potential.
Model 1: Highest stresses occurred at the strut intersections, suggesting that this location may be at highest risk of fracture. This concurs with the in vivo stent fracture data. Model 2: Numerical analyses indicate that the stresses are lower at the strut intersections, but redistributed to the end of the gold reinforcements. This suggests that fractures in this device may occur just distal to the gold. This is indeed the clinical experience. Model 3 was weakest at bolstering the implantation site; however, when two stents were coupled (stent-in-stent technique), better strength and lower stresses were seen compared with Model 1 alone.
Using finite element analysis of known stents, we were able to accurately predict stent fractures in the clinical situation. Furthermore, we have demonstrated that a stent-in-stent technique results in better device performance, which suggests a novel clinical strategy.

Full text

Page 1

C ?2007, the Authors
Journal compilation C ?2007, Blackwell Publishing, Inc.
DOI: 10.1111/j.1540-8183.2007.00294.x
Finite Element Analysis of Stent Deployment: Understanding Stent Fracture
in Percutaneous Pulmonary Valve Implantation
SILVIA SCHIEVANO, M.ENG,1LORENZA PETRINI, PH.D.,2FRANCESCO MIGLIAVACCA, PH.D.,2
LOUISE COATS, M.R.C.P.,1JOHANNES NORDMEYER, M.D.,1PHILIPP LURZ, M.D.,1
SACHIN KHAMBADKONE, M.D.,1ANDREW M. TAYLOR, M.D., M.R.C.P.,
F.R.C.R.,1GABRIELE DUBINI, PH.D.,2and PHILIPP BONHOEFFER, M.D.1
From the1UCL Institute of Child Health and Great Ormond Street Hospital for Children, London, United Kingdom;2Laboratory of Biological
Structure Mechanics, Structural Engineering Department, Politecnico di Milano, Milan, Italy
Objectives: To analyze factors responsible for stent fracture in percutaneous pulmonary valve implantation (PPVI)
by finite element method.
Background: PPVI is an interventional catheter-based technique for treating significant pulmonary valve disease.
Stent fracture is a recognized complication.
Methods: Three different stent models were created: (1) platinum–10% iridium alloy stent – resembles the first-
generation PPVI device; (2) same geometry, but with the addition of gold over the strut intersections – models the
current stent; (3) same design as 1, but made of thinner wire. For Model 3, a stent-in-stent solution was applied.
Numerical analyses of the deployment of these devices were performed to understand the stress distribution and
hence stent fracture potential.
Results: Model 1: Highest stresses occurred at the strut intersections, suggesting that this location may be at
highest risk of fracture. This concurs with the in vivo stent fracture data. Model 2: Numerical analyses indicate
that the stresses are lower at the strut intersections, but redistributed to the end of the gold reinforcements. This
suggests that fractures in this device may occur just distal to the gold. This is indeed the clinical experience. Model
3wasweakestatbolsteringtheimplantationsite;however,whentwostentswerecoupled(stent-in-stenttechnique),
better strength and lower stresses were seen compared with Model 1 alone.
Conclusions: Using finite element analysis of known stents, we were able to accurately predict stent fractures in
the clinical situation. Furthermore, we have demonstrated that a stent-in-stent technique results in better device
performance, which suggests a novel clinical strategy. (J Interven Cardiol 2007;20:546–554)
Introduction
Heart valve disease is generally treated with open
heart surgery. An innovative nonsurgical technique for
heart valve replacement has recently become a real-
ity in the treatment of right ventricular outflow tract
Address for reprints: Silvia Schievano, Cardiothoracic Unit, UCL
Institute of Child Health and Great Ormond Street Hospital for Chil-
dren, Great Ormond Street, London WC1N 3JH, United Kingdom.
Fax: +44 20 7813 8262; e-mail: s.schievano@ich.ucl.ac.uk
Grant: British Heart Foundation FS/05/039.
dysfunction.1,2Percutaneouspulmonaryvalveimplan-
tation (PPVI) involves transcatheter placement of a
valved stent within the existing degenerated valve or
conduit,andprovidesexcellenthemodynamicresults.3
The device is made of a valve from a bovine jugu-
lar vein, sewed into an expandable stent and mounted
on a balloon catheter for delivery (Fig. 1A). Bovine
jugular venous valves are available only up to 22 mm
of diameter. Therefore, only right ventricular outflow
tracts smaller than 22 mm of diameter can be treated
with this percutaneous device. However, in borderline
cases, with larger or high-compliance outflow tracts,
the stent is overdeployed up to 24 mm. The first stent
546Journal of Interventional CardiologyVol. 20, No. 6, 2007

Page 2

STENT FRACTURE IN PERCUTANEOUS VALVE IMPLANTATION
Figure 1. (A) PPVI device: bovine venous valve sutured inside the balloon-expandable stent. Stent fracture in (B) the early
generation device, where the platinum welds between the segments break, and (C) the new design device.
used for PPVI was created by a platinum–10% irid-
ium wire (NuMED Inc., Hopkinton, NY, USA).4The
wirewasformedintoazig-shapedpatternandtheindi-
vidual segments were joined together at the crowns to
create the full stent, by welding of the platinum. Since
the platinum welds were prone to fracture, this device
was modified in design by introducing a gold-brazing
process to reinforce the crowns of the stent.
Between January 2000 and May 2006, PPVI was
successfully performed in 123 patients. The early gen-
erationdevicewasimplantedinthefirst10patients;the
new design prosthesis into the following 113 patients.
Stent fracture is a recognized complication follow-
ing stent implantation for all cardiovascular applica-
tions.5–7In our PPVI series, stent fracture was de-
tected in 26 patients:84/10 (40%) patients treated with
the early generation device and 22/113 (19%) patients
treatedwiththegold-reinforcedstent.Thefracturesoc-
curred during crimping of the stent onto the balloon in
two cases (both early generation devices), following
balloon dilatation in 3 patients, following implantation
ofasecondpercutaneousvalvein1,andspontaneously
in 21 patients. The exact location of fractures in the
PPVI patients was analyzed from both the frontal and
lateral chest X rays. The early generation stents frac-
turedatthestrutintersections(Fig.1B).Afterreinforc-
ing the weld with gold, these fractures were no longer
seen.Infact,inthenewdesignstent,fracturesoccurred
more frequently next to the ends of gold-brazed parts
(Fig. 1C).
In6patients,interventionalmanagementofthestent
fracture was possible by repeat PPVI (stent-in-stent
technique) for stabilization of the fractured parts, with
successful hemodynamic results. The feasibility of
stent-in-stentimplantationhaspreviouslybeendemon-
strated with different stents for a variety of indications
in congenital heart disease.9–11Repeat PPVI repre-
sents the most promising approach to treat stent frac-
ture.Multielementdevicescanbeproducedcombining
stents of diverse material and design to take advantage
of their different mechanical properties, reinforce the
prosthesis, and avoid fracture.
The purpose of this study was to evaluate stent frac-
ture in PPVI devices, by analyzing the stress distribu-
tionduringloadingconditionsandcomparingdifferent
devicedesigns.Finiteelement(FE)analyseswereused
toachievethis.Theextrastrengthprovidedbyasecond
device in the stent-in-stent technique was evaluated in
contrast to the performance of a single prosthesis.
Materials and Methods
Large deformation analyses were performed using
the FE method commercial code ABAQUS/Standard
6.4 (Hibbit, Karlsson & Sorenses, Inc., Pawtucket, RI,
USA).
Geometries and Mesh. Three stent geometries
were created on the basis of data supplied from the
company or obtained from measurements by means
of caliper and optic microscope. The stent geometries
werecreatedtoemulatetheinitialcrimpedstatusofthe
device onto the catheter balloon.
The first model – named PL – was characterized by
six wires (wire diameter of 0.33 mm), each formed in
eight zigzags. Individual wires were joined together at
the crown points to create the full stent (Fig. 2A). This
geometry represented the early generation device used
in PPVI.
The second model (PL-AU) had the same geometry
as the previous one but also included gold-brazed ar-
eas in the shape of 0.076 mm thick sleeves around the
platinum wire crowns (Fig. 2B). This stent resembled
the device currently used.
Vol. 20, No. 6, 2007Journal of Interventional Cardiology547

Page 3

SCHIEVANO, ET AL.
Figure 2. CAD model and dimensions of (A) PL stent and (B) PL-AU stent.
The third model (PL1/2) had the same design as the
PL device but with a wire diameter of 0.23 mm (PL1/2
material mass was half the mass of the PL stent). This
model was designed in order to evaluate and quantify
the change in mechanical performance of a stent made
from a thinner wire.
Thebiologicalvalvemountedintotheclinicallyused
stent was not modeled in this study.
The FE model mesh was automatically generated.
All stents were meshed with 10-node tetrahedrons in
order to easily fit the complex geometries studied and
give an accurate solution. The gold elements of the
PL-AU model were tied to the platinum wires to avoid
relativemovementorseparationbetweenthetwoparts.
Mesh Sensitivity. Before running the analyses, a
sensitivity test was performed on the PL model mesh
to achieve the best compromise between limited calcu-
lation time and no influence of the element number on
the results.
Five meshes with an increasing number of elements
and nodes were tested (Table 1).
Materials
NuMED supplied platinum–10% iridium alloy en-
gineering stress-strain data for uniaxial tension tests
(Young modulus 224 GPa, Poisson ratio 0.37, yield
stress 285 MPa). The material was assumed to have
isotropic properties. A Von Mises plasticity model,
commonly used with metallic alloys, along with an
isotropic hardening law, was used in the analyses.12–14
Handbookpropertieswereappliedforgoldmechanical
behavior (Young modulus 80 GPa, Poisson ratio 0.42,
yield stress 103 MPa).
Analyses. Inflation of balloon-expandable stents is
clinically performed by pressurization of a balloon in-
serted inside the device. Modeling the interaction be-
tween the balloon and the stent is expensive in terms
of time and power calculation and is only important
for analyses in which the transitory configurations are
required.15The intention of this study was to look at
Table 1. Mesh Sensitivity Analysis
SpacingElements NodesRperipheral[%]
A
B
C
D
E
0.17
0.15
0.12
0.115
0.1
85393
95720
166778
218832
284703
176424
195365
324518
417126
527852
1.58
1.57
1.55
1.55
1.54
Rperipehral— elastic recoil in the stent peripheral sections.
548 Journal of Interventional CardiologyVol. 20, No. 6, 2007

Page 4

STENT FRACTURE IN PERCUTANEOUS VALVE IMPLANTATION
Figure3. Deployedconfiguration(black)ofthePL-AUstentresult-
ing from the application of a pressure directly to the internal surface
of the device. The initial configuration of the stent is shown in light
gray.
the stent in its final configuration (when the balloon
was completely inflated) and after balloon deflation.
For these reasons, the balloon was not modeled in the
simulations.
Computationally, the inflation of the stent may be
performedusingeitherdirectpressureappliedtothein-
ternalsurfaceofthestentorthroughprescribedbound-
ary conditions. Attempts to expand the device with a
pressure applied directly to the internal surface of the
stent can prove difficult, due to lack of geometrical
symmetryinthedesign.Indeedtheterminalpartsofthe
stentarenotconstrainedbyothersegments.Therefore,
the deployed configuration could result in unrealistic
deformations of the device (Fig. 3). Consequently, the
stentwasinflatedusingradialexpansiondisplacements
up to an internal diameter of 24 mm (maximum diam-
eter reached by the device during actual PPVI). Once
thestenthadreachedthedesireddiameter,thedisplace-
ment constraints were removed to simulate the balloon
deflation and allow the elastic recoil of the stent.
Lastly, in order to simulate the compression force
experienced by the device due to the implantation site
Figure 4. Three relative rotation degrees between the outer (black) and inner (light gray) devices in the 2PL analyses at the
end of stent inflation.
wall, a gradual pressure was applied to the external
surfaceofthestent.Thisenabledevaluationofthestent
strength to maintain the patency of the vessel.
Tocomparetheperformanceoftwocoupleddevices
(stent-in-stent technique) against the single prosthesis,
the inflation of two stents – one inside the other – was
simulated. First, the outer stent was deployed up to 24
mm and released, as previously described. Next, the
inner device was inflated up to 24 mm, making con-
tact with the outer stent. The displacement constraints
were removed to allow the material to recoil. Finally,
a pressure was applied to the external surface of the
outer stent to evaluate the strength of the structure.
The interaction between the two devices was de-
scribedbyacontactalgorithmwithfriction(coefficient
of sliding friction equal to 0.25).
The stent-in-stent analysis was performed with two
PL(2PL)andtwoPL1/2(2PL1/2)devices.Indeed,three
different coupling configurations of the two PL stents
were analyzed to assess the influence of the relative
position between the inner and outer device: perfectly
aligned (0 degrees), and 11.25 and 22.5 degrees of rel-
ative rotation (Fig. 4). For the PL1/2stent the perfectly
aligned configuration was studied.
Investigated Parameters. The following mechani-
cal properties were measured:
• Elastic recoil (R) following virtual balloon defla-
tion in the stent middle (Rmiddle) and peripheral
(Rperipheral) sections; the elastic recoil was defined
as: R =Dload−Dunload
equal to the stent diameter at the end of the load-
ing and unloading step, respectively. The differ-
enceintheelasticrecoil(?R)betweenperipheral
and middle section of the stent was defined as:
?R = Rperipheral− Rmiddle.
Dload
· 100 , with Dloadand Dunload
Vol. 20, No. 6, 2007Journal of Interventional Cardiology549

Page 5

SCHIEVANO, ET AL.
• Von Mises stress (σVM) map at the end of virtual
balloon inflation, deflation, and after application
of the external pressure.
• Radial strength, represented by the plot of radial
displacement resulting from the applied external
pressure. The displacement was evaluated at both
the peripheral and central nodes of the device.
Results
Mesh Sensitivity. Von Mises stress color map and
elastic recoil of the peripheral nodes of the stent were
checked for the different meshes. The stress distribu-
tion was similar in all meshes. The difference in elastic
recoil between meshes decreased with the increase in
element number (Table 1). Mesh C was selected as the
mesh that guaranteed a solution independent from the
grid without a critical increase in calculation time.
The mesh of the gold parts, built around mesh C of
the PT model, resulted in additional 116,602 elements
for the PL-AU stent.
The PL1/2mesh was made of 149,703 elements and
304,054 nodes.
ElasticRecoil. Inflationbydisplacementcontrolre-
sulted in uniform radial expansion in all stent configu-
rations. Upon balloon deflation, the R of the different
devices was generally low, especially if compared to
the values reported for stents used in different clinical
indications.16–18
Asexpected,RPL1/2> RPLbecauseofthelargerwire
section of the PL stent, and RPL> RPL−AUbecause of
the gold reinforcement in the PL-AU stent (Table 2).
The difference in elastic recoil between the periph-
eral and middle sections was tiny for all the stents. The
highest?RwasinthePL1/2stent,wheretheperipheral
Table 2. Elastic Recoil Values
ModelRperipheral[%]Rmiddle[%]
?R [%]
PL
PL-AU
PL1/2
2PL – 0 degrees
2PL – 11.25 degrees
2PL – 22.5 degrees
2PL1/2– 0 degrees
1.55
1.38
2.31
1.71
1.69
1.70
2.14
1.38
1.16
1.90
1.50
1.52
1.58
1.95
0.17
0.22
0.41
0.21
0.17
0.12
0.19
Rperipheral— elastic recoil in the stent peripheral sections; Rmiddle—
elastic recoil in the stent middle section; ?R— difference in elastic
recoil between peripheral and middle section of the stent.
sections recovered more than the central part. Pressure
applied uniformly to the external surface of the stent
revealedthattheperipheralsectionsofthePL1/2device
were also weaker than the central part in bolstering the
arterial wall (Fig. 5C)
The elastic recoil of the 2PL stent-in-stent analyses
wasalmostthesameinthethreerotationconfigurations
and RPL> R2PL. The coupled system can be imagined
as a combination of two parallel springs. The force
of recovery in the 2PL is bigger than with one single
device.
For the same reason, Rmiddle
Rperipheral
PL1/2
2PL1/2
: the coupling of two PL1/2stents
reinforced the peripheral sections of the structure.
Stress Distribution. The Von Mises stress map at
the inflated diameter of 24 mm is presented in Fig.
5A for the PL, PL-AU, and PL1/2stents. The highest
stressesoccurredinlocalizedregionsofthedevices–at
the strut intersections – where a peak of approximately
660 MPa was detected. These stresses were primarily
due to the bending of the wires close to the platinum
welds as the struts opened during inflation. Stress val-
ues throughout the stent were typically lower, dimin-
ishing rapidly from the crowns to the straight parts.
After virtual deflation of the balloon, at the end of
theelasticrecoil(Fig.5B),σVMwerelowereverywhere
due to the general unloading of the entire structure.
When compared to the PL device, the values of σVM
in PT-AU were slightly smaller, both at the end of the
inflation step (Fig. 5A) and virtual balloon deflation
(Fig. 5B). However, this difference was mostly evident
when the external pressure was applied (Fig. 5C), that
is, when the stent has to resist to the recovering force
of the arterial wall.
The 2PL model gave analogous results in terms of
σVMbetween the three different relative rotation cou-
plings (Figs. 6A, 6B). The stress distribution in the
inner 2PL stent was similar to that of the PL stent.
However, the outer 2PL stent presented lower stress
valuesthanthePLdeviceduringtheentireloadinghis-
tory. The same results were found for the 2PL1/2inner
and outer stents (Fig. 6C) when compared to the PL1/2
model.
Device Strength. The charts in Figure 7 show the
radial displacement of the peripheral and middle sec-
tion nodes of the stents subject to the external pres-
sure. The trend lines were similar in the two sections
for all devices: at low pressure levels, high increases
in pressure corresponded to low displacements, as the
devices possessed adequate strength. However, as the
PL1/2< Rmiddle
2PL1/2. However,
> Rperipheral
550Journal of Interventional Cardiology Vol. 20, No. 6, 2007

Page 6

STENT FRACTURE IN PERCUTANEOUS VALVE IMPLANTATION
Figure 5. Von Mises stress (σVM) map in the PL, PL-AU (the gold elements were removed to visualize the stress distribution
in the platinum elements), and PL1/2stents at (A) end of the inflation, (B) elastic recoil, and (C) after application of a 0.2
MPa pressure to the external surface of the devices.
pressure increased past a threshold, all structures lost
their strength, and displacement increased dispropor-
tionately to pressure. The threshold pressure for each
type of stent was different depending on its design.
As expected, the weaker device was the PL1/2stent,
because of the thinner wire used to form it. The gold
brazing reinforced the PL stent providing it with extra
strength. The relative rotation between the inner and
Vol. 20, No. 6, 2007Journal of Interventional Cardiology551

Page 7

SCHIEVANO, ET AL.
Figure 6. Von Mises stress (σVM) map of the inner and outer stents
of the 2PL models for (A) 0 degrees and (B) 22.5 degrees configu-
ration, and (C) of the 2PL1/2model, at 0.2 MPa of pressure.
outer stent in the 2PL devices did not influence the
displacementresponsetotheappliedpressure.The2PL
model presented a higher strength than the single PL
deviceandeventhanthePL-AUstent,especiallyinthe
peripheral sections. The 2PL1/2device was stronger
than the single PL1/2and its strength was comparable
to the PL stent.
Discussion
The targets involved in the design of a PPVI stent
require a careful compromise between interrelated
and sometimes contradictory material and geometri-
cal properties. The PPVI stent must be loaded onto
the delivery system and manually crimped on the bal-
loon.Uponinsertionintothevascularenvironment,the
delivery system must be manipulated within the tortu-
ous anatomical pathways leading to the implantation
site. The delivery of the device to the optimal posi-
tion requires good visibility under fluoroscopy. Stent
deployment is gained by gradual inflation of the bal-
loon. Upon acquiring the final diameter, the balloon
deflation causes recoil of the stent to a smaller diam-
eter, which is also influenced by the pressure exerted
by the implantation site wall. High stent recoil rate can
causestentdislodgment.Stentstructuralintegritymust
be guaranteed in the long term. The major concerns re-
lated to fracture are the maintenance of radial strength,
integrity of the sutures between valve and stent, and
the risk of late embolization in view of the lack of tis-
sue ingrowth that we have seen with the PPVI stent.
A high biocompatibility is also necessary to prevent
thromboses or restenoses. During follow-up, standard
X-ray-based imaging investigation has to be used to
assess the performance of the device.
Although platinum and iridium are mechanically
ratherweakmaterials,theyalsopresentsomeimportant
characteristics that make them the materials of choice
for the PPVI stent. Platinum–10% iridium alloy is bio-
compatible and has an exceptional radio-opacity due
to its high density (21.55 g/cm3against 7.95 g/cm3of
stainless steel). The resulting high radio-visibility per-
mitstheuseofthinwires,thusimprovingflexibilityand
deliverability. Indeed, the stent wire diameter has to be
as small as possible to have the minimum overall pro-
file to negotiate the vascular pathways. The use of this
material and the unique wire-based design facilitates
crimping onto the balloon and allows stent expansion
at acceptable balloon pressures. The reasonably small
elastic recoil (<2%) guarantees a safe anchoring of the
device in the implantation site. The breakage of the
platinum welds at the crown junctions has been solved
by gold brazing. However, the resistance to fracture of
the new PPVI device remains the major concern of this
stent material and design.
This FE study has proved that the maximum
stresses reached in the device during inflation remain
552Journal of Interventional CardiologyVol. 20, No. 6, 2007

Page 8

STENT FRACTURE IN PERCUTANEOUS VALVE IMPLANTATION
Figure7. Radialdisplacementofthestentperipheralandmiddlenodesinresponsetotheexternalpressureappliedtoemulate
the compression force of the implantation site.
acceptable (platinum–10% iridium ultimate tensile
strength of 875 MPa, data supplied by the manufac-
turer). However, it is clearly visible from the computa-
tionalanalysesthatthestressincreasesaccordingtothe
expansion diameter; the safety of the device, therefore,
is highly dependent on the magnitude of deployment.
The comparison between the PL and PL-AU models
after external pressure application showed much lower
stressinthePL-AUstentatthestrutintersections.This
is because in these points the resistant section of the
PL-AU device is larger. The relatively weak gold ac-
tually reinforces the weld sections of the stent. How-
ever, it is possible to note a redistribution of σVMin
the straight platinum sections, at the end of the gold
reinforcements: the structure is loaded at these points
more than when there is no reinforcement, because of
thereinforcementitself.Indeed,thegoldreinforcement
creates geometrical and material discontinuities. This
suggests that fractures in the PL-AU device may occur
just distal to the gold-brazed elements, as proven from
patient X-ray investigation (Fig. 1C).
The limited recent experience with the stent-in-stent
technique demonstrates not only that repeat PPVI is
safe and feasible, but also that the implantation of a
previous device before the valved one may act func-
tionally to bolster the vessel and ensure the integrity
of the valved stent. The 2PL1/2device compared to the
PL stent showed the same ability to withstand external
pressure, the same stress distribution in the inner stent,
but favorable, lower stress values in the outer device.
Because of its wire diameter, the two PL1/2stents em-
ployed in the 2PL1/2model present the same material
mass as the PL stent, but the thinner wire allows easier
crimping, better deliverability, and greater flexibility.
The recoil is higher in the 2PL1/2device than the PL
stent. However, the FE study showed that as gold braz-
ing reinforces the platinum wires, the elastic recoil is
reduced. Therefore, it is reasonable to conclude that a
coupling of two PL-AU devices made of a thinner wire
wouldresultinbetterperformance.Inclinicalpractice,
the stent-in-stent technique could be performed either
sequentiallyorasasingle-stepapproach.Inourexperi-
ence,sequentialstentingcanbeperformedsafely(6pa-
tients – no embolization). For the single-step approach
(not performed to date), the delivery system would be
of similar size, as the two stents are half the size of
the previous single device, and thus additional com-
plications of a higher French delivery system would be
avoided.Therefore,themultielementstentwithgoldre-
inforcement is a possible solution to reduce the chance
of device fracture in PPVI and increase the success of
this procedure, without theoretically compromising its
technical ease. The FE analysis of this device was not
carried out, because of the large number of nodes and
elements required by the mesh of this model, which
exceeds the performance of the computer used in this
study.
The pressure to compress the stents modeled in this
study to a smaller diameter (Fig. 7) is high if compared
to data reported from mechanical tests in endovascu-
lar stents,19,20where, however, the device is subject to
punctual loads. The pressure in the FE model is uni-
formly applied along the stent circumference. There-
fore,higherpressurevaluesareguaranteedtobeonthe
safe side in the computational analysis.
In vivo, the stent conforms its shape to the implanta-
tion site.8Some stent dimensions were assessed from
angiographic pictures in the PPVI patients. The mea-
surements showed that the shape of the in vivo stent
differs from the theoretical cylindrical profile. There-
fore, the forces that the stent may be subjected to by
the implantation site and the surrounding tissues are
not uniform around the circumference. This can cause
high-stress concentrations in some parts of the stents
Vol. 20, No. 6, 2007Journal of Interventional Cardiology 553

Page 9

SCHIEVANO, ET AL.
and increase the risk of fracture. Clearly, the main lim-
itation of this study is the absence of the implantation
site model.21,22The next step will be the development
of FE models of realistic right ventricular outflow tract
geometries, which can be obtained from intravascu-
lar ultrasound or magnetic resonance imaging.23,24By
simulatingtheinteractionbetweenthestentandthereal
implantation site model, it may be possible to evaluate
the deformed shape of the device and the real distribu-
tion of stresses to which the prosthesis is subjected.
Conclusions
PPVIisasuccessfulalternativetoopenheartsurgery
forpulmonaryvalvediseasetreatment.Stentfractureis
the major complication related to this procedure. Few
in vivo and experimental data about the factors respon-
sible for PPVI stent fracture have been available until
now.Inthispaper,theFEmethodisproposedasatech-
nique to analyze and compare existing stents in order
to understand the mechanical reasons for their fracture
and to optimize their design. The multielement stent
presentedinthisstudyisanewconceptthatcouldsolve
the problem of fracture in PPVI devices. The coupling
of two stents, made from thin wires, results in high
strength and low stresses, which guarantees better re-
sistance to fracture, without affecting other fundamen-
tal device properties such as easy crimping and low
elastic recoil.
Acknowledgments:. The authors would like to thank Dou-
glasVillnave,AllenTower,andTimRyanfortheirtechnicalsupport.
References
1. Bonhoeffer P, Boudjemline Y, Saliba Z, et al. Percuta-
neous replacement of pulmonary valve in a right-ventricle to
pulmonary-artery prosthetic conduit with valve dysfunction.
Lancet 2000;356:1403–1405.
2. Bonhoeffer P, Boudjemline Y, Qureshi SA, et al. Percuta-
neous insertion of the pulmonary valve. J Am Coll Cardiol
2002;39:1664–1669.
3. Coats L, Khambadkone S, Derrick G, et al. Physiological and
clinical consequences of relief of right ventricular outflow tract
obstruction late after repair of congenital heart defects. Circu-
lation 2006;113:2037–2044.
4. Cheatham JP. Stenting of coarctation of the aorta. Catheter
Cardiovasc Interv 2001;54:112–125.
5. Murugan SJ, Dickinson DF, Gibbs JL. Images in cardiology.
Completefractureofarightventricletopulmonaryarteryshunt.
Heart 2005;91:1441.
6. Peng LF, McElhinney DB, Nugent AW, et al. Endovascu-
lar stenting of obstructed right ventricle-to-pulmonary artery
conduits – a 15-year experience. Circulation 2006;113:2598–
2605.
7. Scheinert D, Scheinert S, Sax J, et al. Prevalence and clinical
impact of stent fractures after femoropopliteal stenting. J Am
Coll Cardiol 2005;45:312–315.
8. Nordmeyer J, Khambadkone S, Coats L, et al. Risk stratifi-
cation, systematic classification and anticipatory management
strategiesforstentfractureafterpercutaneouspulmonaryvalve
implantation. Circulation 2007;115:1392–1397.
9. Powell AJ, Lock JE, Keane JF, et al. Prolongation of RV-PA
conduit life span by percutaneous stent implantation. Interme-
diate term results. Circulation 1995;92:3282–3288.
10. Knirsch W, Haas NA, Lewin MAG, et al. Longitudinal stent
fracture 11 months after implantation in the left pulmonary
arteryandsuccessfulmanagementbyastent-in-stentmaneuver.
Catheter Cardiovasc Interv 2003;58:116–118.
11. Sugiyama H, Williams W, Benson LN. Implantation of en-
dovascular stents for the obstructive right ventricular outflow
tract. Heart 2005;91:1058–1063.
12. Dumoulin C, Cochelin B. Mechanical behaviour modelling
of balloon-expandable stents. J Biomech 2000;33:1461–
1470.
13. Poncin P, Proft J. Stent tubing: Understanding the desired at-
tributes. Proceedings of the Materials & Processes for Medical
Devices Conference, Anaheim (CA), USA, 2003.
14. Petrini L, Migliavacca F, Auricchio F, et al. Numerical investi-
gation of the intravascular coronary stent flexibility. J Biomech
2004;37:495–501.
15. Wang WQ, Liang DK, Yang DZ, et al. Analysis of the transient
expansion behavior and design optimization of coronary stents
by finite element method. J Biomech 2006;39:21–32.
16. Ratib K, Doshi SN, Townend JN. Double-stenting of an os-
tial left main-stem lesion for elastic recoil. Int J Cardiol
2005;102:357–358.
17. Schmidt W, Andresen R, Behrens P, et al. Characteristic me-
chanical properties of balloon-expandable peripheral stent sys-
tems. Rofo 2002;174:1430–1437.
18. Grenacher L, Deutsch J, Lubienski A, et al. Resistance to hoop
stress in balloon expandable stents: Evaluation in an Ex vivo
model. Invest Radiol 2003;38:65–72.
19. Rieu R, Barragan P, Masson C, et al. Radial force of coro-
narystents:Acomparativeanalysis.CatheterCardiovascInterv
1999;46:380–391.
20. Dyet JF, Watts WG, Ettles DF, et al. Mechanical proper-
ties of metallic stents: How do these properties influence the
choice of stent for specific lesions? Cardiovasc Intervent Ra-
diol 2000;23:47–54.
21. MigliavaccaF,PetriniL,MassarottiP,etal.Stainlessandshape
memoryalloycoronarystents:Acomputationalstudyonthein-
teraction with the vascular wall. Biomech Model Mechanobiol
2004;2:205–217.
22. Lally C, Dolan F, Prendergast PJ. Cardiovascular stent de-
sign and vessel stresses: A finite element analysis. J Biomech
2005;38:1574–1581.
23. Schievano S, Coats L, Migliavacca F, et al. Variations in right
ventricular outflow tract morphology following repair of con-
genitalheartdisease–implicationsforpercutaneouspulmonary
valve implantation. J Cardiovasc Magn Reson 2007;9:687–
695.
24. SchievanoS,MigliavaccaF,CoatsL,etal.Planningofpercuta-
neouspulmonaryvalveimplantationbasedonrapidprototyping
of the right ventricular outflow tract and pulmonary trunk from
magnetic resonance imaging data. Radiology 2007;242:490–
497.
554Journal of Interventional Cardiology Vol. 20, No. 6, 2007