959

Functional and structural effects of percutaneous

edge-to-edge double-orifice repair under cardiac
cycle in comparison with suture repair
A Avanzini*, G Donzella, and L Libretti
Department of Mechanical and Industrial Engineering, University of Brescia, Italy

The manuscript was received on 14 February 2011 and was accepted after revision for publication on 3 June 2011.

DOI: 10.1177/0954411911414803

Abstract: Percutaneous procedures for double-orifice mitral valve repair using the MitraClipÒ
device (clip) have been recently introduced as new treatment options as alternatives to medi-
cal management and open-heart surgery, especially for patients with high estimated operative
risk. Similarly to the open-heart surgical technique, where suturing is used, the clip creates a
double-orifice configuration that not only improves the closing function of the valve, but also
significantly modifies its behaviour, particularly in the diastolic phase. While several clinical
trials have been conducted, and are ongoing, in order to assess the safety and effectiveness of
this technique, a deeper knowledge of the structural and functional effects on the valve, and
of the cyclic loads transmitted to the clip itself, would allow a comparison with other repair
techniques, and could serve as a foundation for possible further optimization of the clip
design. The effects of the MitraClipÒ device developed by Evalve Inc. were studied by means
of a finite element model of the mitral valve, specifically developed to study the structural
effects of the original, suture-based, edge-to-edge technique. A second model was developed
in order to simulate the effects of a suture with similar extension from the leaflet edge in a
direction to the annulus, in order to compare the two repair techniques. The mitral valve area
and transvalvular pressure gradient predicted by the models for the clip and the suture are
quite similar. Similar leaflet cyclic stresses, both in value and in location, were noted for the
two mechanisms of linking the leaflets, while minor differences were found in the load trans-
mitted to the suture and the clip, with slightly higher values for the clip. The model satisfacto-
rily allowed functional parameters (valve area and transvalvular pressure gradient) and
structural parameters (load, leaflet stress) to be determined. Overall, the structural effects of
the clip and the suture are quite similar under the cyclic loading conditions imposed by the
cardiac cycle.

Keywords: double orifice, MitraClip, percutaneous repair

1 INTRODUCTION enhancing the quality of life and long-term survival

Since 1991, the surgical double-orifice or edge-to-
edge repair technique has been used in patients with
leaflet prolapse to treat mitral valve regurgitation,
*Corresponding author: Department of Mechanical and
Industrial Engineering, University of Brescia, via Branze,
38-25123 Brescia, Italy.
email: andrea.avanzini@ing.unibs.it
for thousands of patients [1, 2]. The technique allows
repairs in more complex valve pathology and avoids
of valve replacement, and may thereby avoid compli-
cations such as thromboembolism, and the need for
long-term anticoagulation treatment, associated with
valve replacement. Mitral repair preserves the papil-
lary muscles and chordae, and therefore left ventricu-
lar function, and is associated with lower operative
and late mortality [3, 4]. Contraindications to surgery
do exist in patients with high estimated operative risk

Proc. IMechE Vol. 225 Part H: J. Engineering in Medicine

Downloaded from pih.sagepub.com at The University of Iowa Libraries on June 19, 2015
960 A Avanzini, G Donzella, and L Libretti

and multiple co-morbidities. In order to overcome obtained with a standard edge-to-edge suture tech-
these limitations, percutaneous procedures for heart nique during diastole and systole.
valve disease have recently been introduced [5].
Because of their minimally invasive nature, they are 2 METHOD
associated with a low risk of procedure-related com-
plications, but optimal patient selection and long-
The effects of the clip were studied by means of a
term effectiveness have still to be determined [6].
hyperelastic finite element model of the mitral valve
Regarding percutaneous mitral leaflet repair, one
specifically developed to study the structural effects
of the most promising approaches has been devel-
of the surgical technique. Two different models
oped by Evalve Inc. (Abbott Vascular, Santa Clara,
were implemented, in order to simulate the pres-
CA) using the MitraClipÒ system [7]. It consists of a
ence on the same mitral valve either of the clip,
MitraClipÒ device and a clip delivery system, which
modelled as an independent part with dimensions
is advanced through a steerable guide catheter. The
corresponding to its actual size, or of a suture with
procedure consists in positioning the clip following
similar extension along the leaflet edge. In the fol-
standard trans-septal access to the left atrium under
lowing, these models will be referred to as model 1
echocardiographic guidance, and joining together a
(MitraClipÒ device, with width 3.5 mm and length
portion of the anterior and posterior leaflets at the
7 mm) and model 2 (suture, with width 4 mm and
location of mal-coaptation.
length approximately 2.5 mm).
Several clinical trials [8] have been performed,
After valve geometry definition, structural analy-
and are ongoing, in order to assess the safety and
ses were carried out by means of the commercial
effectiveness of this technique in comparison with
software ABAQUS 6.8.1 (SimuliaÓ). The program
standard cardiac surgery. Up to now, the results
used for finite element modelling of the mitral valve
have demonstrated that percutaneous clip repair is
is a specialized program for structural analysis, par-
able to reduce mitral regurgitation, with a low mor-
ticularly suitable for the study of complex and
bidity and mortality rate and meaningful clinical
sophisticated models of biological apparatus.
benefits, in patients with both degenerative and
functional mitral regurgitation.
2.1 Mitral valve geometry
Similarly to the double-orifice repair with suture,
the presence of the clip creates a double-orifice con- The mitral valve apparatus mainly consists of two
figuration that can significantly modify the beha- valve leaflets, anterior and posterior, connected
viour of the valve, particularly in the diastolic phase. along the valve perimeter to a fibrous ring called the
It is therefore essential to assess the stress on the annulus. On the surface facing the ventricle, leaflets
leaflets induced by the clip, and the loads sustained are connected to papillary muscles by marginal and
by the clip during the cardiac cycle. A deeper knowl- strut chordae tendinae.
edge of the structural effects of this procedure on A scheme of the valve model adopted in the pres-
the valve, and of the loads transmitted to the clip ent work is represented in Fig. 1.
itself, would allow a comparison with other repair The complex anatomical structure of the mitral
techniques, and could serve as a basis for possible apparatus was represented by deriving the valve
optimization of clip design. Very limited data are
available in the literature concerning these topics,
and none are available referring to the application
of a clip. Only a few numerical studies that consid-
ered the presence of a surgical suture [9, 10], and
experimental measurements carried out in vivo and
in vitro with miniature load transducers joining the
leaflets [11–13], are available.
In order to fill this gap in the present knowledge,
a finite element model was developed to simulate
the clip and to calculate the corresponding trans-
mitted load and leaflet stresses during diastole and
systole. The functional parameters (mitral valve area
and transvalvular pressure gradient) and the struc-
tural effects (load transmitted to the clip and stress
state in the leaflets) due to the application of the
clip were then determined and compared with those Fig. 1 Mitral valve: atrial view

Proc. IMechE Vol. 225 Part H: J. Engineering in Medicine

Downloaded from pih.sagepub.com at The University of Iowa Libraries on June 19, 2015
 Effects of percutaneous edge-to-edge double-orifice repair under cardiac cycle in comparison with suture repair
geometry directly from the model proposed in refer-
ence [9]. In that work the annulus size was 5 cm2
(corresponding to a largest diameter of about
30 mm). This is a value in the range reported in the
literature for a healthy mitral valve; it also corre-
sponds to the size of the prosthetic ring usually
implanted during repair intervention to restore phy-
siological valve dimension. However, the current clip
repair procedure does not include annuloplasty, and
furthermore it is well known that many valves with
degenerative disease may present a significantly
enlarged MVA compared with their original dimen-
sions. For this reason both model 1 and model 2 used
in the present study, which share the same valve geo-
metry, were created by a scaling operation in which
the annulus size was enlarged to 10 cm2 (correspond-
ing to a largest diameter of about 43 mm). In order to
keep the same ratios as in the original model
between the anterior-posterior diameter (B) and the
commissure-to-commissure diameter (A), and
between the anterior and posterior leaflet length, a
scaling factor of approximately 1.42 was assumed for
these characteristic dimensions (see Table 1).
It should be noted that the scaling operation
starting from a smaller valve does not really corre-
spond to the changeover from a healthy to a patho-
logical valve; it was used only for ease of creation of
the desired geometry starting from an existing
model, taken as a reference.
Table 1 Mitral valve dimensions
Model Annulus A (mm) B (mm)
area (cm2)
Present 10 42.74 28.14
Reference [9] 5 30 19.75
Fig. 2 (a) MitraClipÒ device; (b) model
Downloaded from pih.sagepub.com at The University of Iowa Libraries on June 19, 2015
961
The leaflets’ thickness was kept the same as in
the previous model (1.32 mm for the anterior leaflet
and 1.26 mm for the posterior leaflet [14]).
The model includes some simplifying assump-
tions concerning the valve geometry (i.e., absence of
clefts in the posterior leaflet, physical separation of
anterior and posterior leaflet in the commissural
region), but it still correctly reproduces the main
features of the mitral apparatus, and at the same
time it guarantees computational efficiency and
accuracy, as shown by previous studies.
Both suture and clip models implicitly take into
account possible leaflets prolapse, as the repair
simulation introduces a local constraint (described
in the following) if and where prolapse is present. In
addition, compared with the values reported in liter-
ature for a reference mitral valve [15, 16], the geo-
metry considered here refers to a valve having a
higher annulus area (10 cm2 rather than 8 cm2) but
similar leaflet lengths (about 24 mm and 14 mm),
thus also addressing the condition of pure annulus
dilatation. As a consequence, the model fits patients
with both pure annular dilatation and ruptured
chordae.
2.2 MitraClipÒ device and suture schemes
The clip (shown in Fig. 2(a)) is a complex mechani-
cal device consisting of many different components.
Anterior leaflet Posterior leaflet Annulus
length (mm) length (mm) perimeter (mm)
23.9 14.3 116.95
16.8 10 81.6
Proc. IMechE Vol. 225 Part H: J. Engineering in Medicine
962 A Avanzini, G Donzella, and L Libretti
In the present study its actual geometry has been
simplified by considering the clip itself as consisting
of the two arms that perform coaptation of the leaf-
lets. The clip was modelled as consisting of
two symmetrical arms connected by a hinge in its
bottom part (Fig. 2(b)). When the device is closed,
the arms are almost parallel, and the distance
between them is equal to the sum of the leaflets
thickness. The upper parts of the arms are inclined
at a small angle to allow a slight separation between
the leaflets, reproducing their V-shape. A radius is
present at the top part of the arms, to allow a
smooth transition from the area constrained by the
clip to the free surfaces of the leaflets. The arm
thickness assumed for the clip is 1 mm.
The suture has been simulated using link connec-
tors supported by coupling constraints for each of
the two leaflets, with the function of distributing the
transmitted load on an area corresponding to the
extension of the actual Teflon pledgets.
When comparing the constraint imposed by the
two repairs in the present model, it should be noted
that the grasp length is not exactly the same for the
two techniques, and moreover that, in its upper
part, the clip exerts a unilateral constraint (from the
outside) on the leaflets’ surfaces, which is not pres-
ent in the case of the suture.
2.3 Finite element models
Based on the mitral valve and clip geometry, the finite
element models described below were developed.
2.3.1 Materials
Mitral valve leaflets consist of a soft biological tis-
sue, essentially made of a fibrous tissue network,
mainly collagen and elastin, saturated with a fluid
that is mostly water.
As a result of the coupling between different con-
stituents, the typical stress–strain diagram may be
divided into a pre-transition region, in which the
stress strain curve is fairly flat, followed by a post-
transition region where the slope suddenly increases.
Many different theoretical approaches have been
proposed for constitutive modelling of this complex
mechanical behaviour. In the present work a hypere-
lastic material constitutive law has been adopted, in
which the strain-energy potential U has been
defined in the reduced polynomial form as
X
N  i XN
1  el 2i
U= Ci0 I1 3 + J 1
i=1
D
i=1 i
where I1 is the first deviatoric strain invariant; Jel is
the elastic volume ratio; N is the order of the
Proc. IMechE Vol. 225 Part H: J. Engineering in Medicine
Downloaded from pih.sagepub.com at The University of Iowa Libraries on June 19, 2015
Table 2 Material coefficient (units MPa)
C10 C20 C30 C40 C50
8.733 3 1022 21.236 12.912 250.532 95.518
polynomial function, taken equal to 5; and Ci0
and Di0 are material parameters, determined by a
simple fitting procedure, automatically available in
the Abaqus code, and described in more detail in
reference [9].
The values of the coefficients Ci0 are reported in
Table 2. Under the assumption of material incom-
pressibility, the coefficients Di were all set equal
to zero.
Once the strain-energy function is known, the
stress state can be determined by taking the deriva-
tive of U with respect to a strain measure
∂U
s= (2)
∂e
where s and e are any work conjugate stress and
strain measures. As previously mentioned, the leaf-
let material consists mainly of water and therefore
has very little compressibility. Incompressibility is
in general a difficult phenomenon to handle com-
putationally; nevertheless, when using shell ele-
ments in applications where the material is not
highly confined, it is quite satisfactory to assume
that the material is fully incompressible. The clip is
made mainly of metallic materials whose stiffness is
greater than that of the leaflets by orders of magni-
tude. Since the stress state in the device is not con-
sidered in these simulations, its material is simply
assumed as infinitely rigid by assigning to the clip a
value of Young’s modulus that is extremely high
compared with the leaflet tissues stiffness.
2.3.2 Contact interactions and boundary conditions
Valve leaflets are involved in a different kind of con-
tact interaction, whose correct definition is funda-
mental. Leaflet coaptation at valve closure during
systole has been modelled considering frictionless
contact between their atrial surfaces. The clip is
considered as an independent geometric part whose
structure interacts with the leaflets. Contact
between the leaflets’ ventricular surfaces and clip
arms has been considered as rough – that is, no
separation or relative sliding is allowed after contact
(1)
is reached – for better simulation of the device’s
strong fixing action. In order to model precisely the
contact interaction in the clipping region and the
real kinematic constraints imposed by the clip, it
 Effects of percutaneous edge-to-edge double-orifice repair under cardiac cycle in comparison with suture repair
Fig. 3 Boundary conditions: (a) fixed annulus; (b) schematization of chordae tendinae constraint
was decided to adopt a surface-to-surface approach,
taking as reference surfaces for contact pairs the
inner surfaces of the clip, which actually interacts
with the valve leaflets, and the midplane surfaces of
the leaflets. The boundary conditions reported in
Fig. 3 were initially applied; in particular, a fixed
constraint (no displacement) was imposed on the
nodes of the fibrous annulus, and the effect of
the chordae tendinae was simulated by constraining
the free-edge nodes of the leaflets to move on a plane
inclined at 26.5° with respect to the annulus plane.
This boundary condition allows correct simulation of
the constraint imposed by the chordae tendinae on
leaflet movement without physically modelling this
apparatus, thus significantly reducing the model
complexity and the related convergence problem,
especially for the contact interaction phases.
2.3.3 Mesh
The leaflet mesh consists of three- and four-node
reduced-integration shell elements with hourglass
control and finite membranal strain. In order to
obtain satisfactory convergence in all configurations
examined, slightly different meshes were used for
the clip and suture models. In particular, for model
1 (the clip), the anterior and posterior leaflets were
meshed with 6736 and 3550 elements, respectively.
Each clip arm consists of 792 elements, for a total of
1584. In order to calculate the load transmitted to
the clip, a link connector was introduced in the cen-
tre of the device. For model 2 (the suture) the ante-
rior and posterior leaflets were meshed with 5506
and 3550 elements, respectively. The suture was
Downloaded from pih.sagepub.com at The University of Iowa Libraries on June 19, 2015
963
modelled using rigid connector elements (link type)
to simulate suture stitches. In particular, the suture
was simulated with six link connectors (for a total
width of 4 mm) coupled to a small surface near the
leaflet edges. In this case, the total force transmitted
to the suture is given by the algebraic sum of those
of each link connector.
2.4 Analyses
In order to simulate the surgical procedure and the
effect of clip positioning, FEM analyses were carried
out considering two main steps: the first included
leaflets approach and clip or suture application, and
the second involved pressure application to simu-
late the cardiac cycle in diastole and systole.
Different phases of the FEM analyses, which are
slightly different depending on the presence of the
clip or of the 4 mm suture, are detailed below and in
Figs 4–6, in which, for convenience, only half the
mitral valve is displayed.
2.4.1 Leaflets approach and MitraClipÒ device
or suture application
Model 1 - MitraClipÒ device. The clip application is
described below (see Fig. 4).
(a) The hinge of the clip is positioned in corre-
spondence with the leaflet border in the middle
of the valve. The clip is completely open, in a
position parallel with respect to the orifice
plane; the hinge is constrained to a fixed posi-
tion. A pressure of about 30 per cent of systolic
Proc. IMechE Vol. 225 Part H: J. Engineering in Medicine
964 A Avanzini, G Donzella, and L Libretti

Fig. 4 Clip application steps: (a) clip positioning; (b) arms closure; (c) clip locking; (d) catheter removal

Fig. 5 Complete model of mitral valve after clip application

Fig. 6 Suture application phases: (a) leaflets approach; (b) suture positioning

is applied on the ventricular side to bring the (b) Closure of the clip is simulated by rotating its
leaflets closer so to allow successive clip arms around the fixed hinge; at the end of the
application. closure its arms present an angle of about 15°

Proc. IMechE Vol. 225 Part H: J. Engineering in Medicine

Downloaded from pih.sagepub.com at The University of Iowa Libraries on June 19, 2015
 Effects of percutaneous edge-to-edge double-orifice repair under cardiac cycle in comparison with suture repair
with respect to the vertical plane. The central
link, initially deactivated, is activated at the end
of the closure, simulating the clip locking.
(c) The pressure applied on the leaflet surfaces
during phase (a) is removed. At this point the
regions of the leaflet surfaces that came into
contact with clip arms are constrained to the
clip itself (as a result of rough type interaction).
(d) The fixed position constraint imposed on the
hinge is removed. The device is then free to
move in the space, remaining only in contact
interaction with the leaflets. At the end of this
phase, which corresponds to the catheter being
removed, the valve–clip system reaches the
equilibrium configuration.
A complete view of the valve model with the clip
after the application phases is shown in Fig. 5.
Model 2: Suture 4 mm width. At the beginning of
the analysis, the same boundary conditions as for
the clip are applied; then the following sequence is
applied (see Fig. 6).
(a) A pressure of about 10 per cent of systolic is
applied on the ventricular side to bring the
leaflets closer. The applied pressure is smaller
than that necessary for clip application,
because a large area of leaflet coaptation is not
required, the suture being created very close to
the leaflet edge.
(b) The links that simulate the suture are activated
in the model, and the pressure is deactivated.
The leaflets reach the equilibrium position
without loads.
It should be noted that the initial application of a
pressure on the ventricular side is only a modelling
expedient for bringing the leaflets together, thus
simulating the technique used by the surgeon in
order to proceed with the suture. The extent of this
‘virtual’ pressure can change, depending on valve
geometry, but it does not enter in the simulation of
the cardiac cycle, because it is removed once the
repair is applied.
2.4.2 Simulation of the cardiac cycle
Diastolic phase. After either clip or suture application,
the chordae tendinae constraint is removed for both
models, on the assumption that the chordae have no
structural effects during diastole, and an increasing
pressure is applied on the atrial side of the valve. In
this phase, the leaflets open in a double-orifice con-
figuration. The mitral valve area (MVA) and transvalv-
ular pressure gradient (Dp) in the diastolic phase
Downloaded from pih.sagepub.com at The University of Iowa Libraries on June 19, 2015
965
were determined using the procedure described in
reference [17], based on the fluid-dynamic FEM
model proposed in reference [18].
In particular, under the hypothesis of a constant
blood flow rate across the valve, Q (l/min), Dp
(mmHg) was determined as a function of MVA
(cm2) by means of the equation
 2
Q
Dp = 0:16 (3)
MVA
This equation represents all possible working
points of the valve. Following the literature [19], a
typical value of Q = 13.7 l/min has been assumed in
present analyses. By computing the MVA from the
valve FEM model at different pressure levels, it is
possible to identify the valve’s effective working
point.
Systolic phase. After the clip or suture application,
pressure from the ventricular side is applied, main-
taining the chordae tendinae constraint. The
applied systolic pressure has a value of 130 mmHg,
in the physiological range reported in the literature.
3 RESULTS
3.1 Model 1: MitraClipÒ device
3.1.1 Diastolic phase
According to the procedure previously described,
the predicted MVA for the valve repaired with the
clip is 1.83 cm2, and the transvalvular pressure gra-
dient is 8.94 mmHg.
The tensile force transmitted to the link in dia-
stole reaches a value of 0.39 N, increasing almost
linearly with applied pressure.
The Von Mises stress contour map of the leaflets
under the effect of diastolic pressure is reported in
Fig 7. From this figure, it can be observed that stress
peaks are located near the top edges of the clip
arms. The points where the stress is maximum are
indicated as AC in the figure.
3.1.2 Systolic phase
Because of the systolic pressure application, a com-
pressive (closing) force is transmitted to the clip
arms with a peak of 0.43 N, whose existence pre-
sumes a perfect sealing between clip arms and leaf-
lets. The Von Mises stress contour map of the
leaflets under the effect of systolic pressure is
reported in Fig 8. In this case, a large, highly
stressed zone can be observed in two symmetrical
central parts of the anterior leaflet, in addition to
Proc. IMechE Vol. 225 Part H: J. Engineering in Medicine
966 A Avanzini, G Donzella, and L Libretti

Fig. 7 Diastole with MitraClipÒ device: leaflets Von Mises stress distribution (MPa)

Fig. 8 Systole with MitraClipÒ device: leaflets Von Mises stress distribution (MPa)

the local effect generated by the clip. The maximum
stress occurs in these central parts, at the point indi-
cated as BC in the figure.
In Table 3, the in-plane principal stress values sI
and sII are reported for points AC and BC defined
above, on both the atrial and ventricular sides of the
leaflets, at the end of diastole, and at the end of sys-
tole. The corresponding stress ranges during the
cardiac cycle (i.e. the difference between the diasto-
lic and systolic values) are reported along with the
equivalent alternating Von Mises stress, calculated
as
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
saMises = s2Ia + s2IIa sIa sIIa
where sIa and sIIa are the alternating principal
stress components, i.e. one half of the correspon-
dent stress ranges.
The equivalent alternating Von Mises stress is a
parameter suitable for evaluating the structural
effects due to fatigue on the leaflet tissues. The
resulting values may be useful in comparing the
predicted cyclic damage level in the highly stressed
points highlighted above, and for the two repair
techniques examined.
From the above results, the presence of both
membranal and flexural stresses on the leaflets dur-
ing the cardiac cycle can be observed. It should also
be noted that the highest stressed zone is point BC,
(4) on the atrial side of the leaflet body during systole,

Proc. IMechE Vol. 225 Part H: J. Engineering in Medicine

Downloaded from pih.sagepub.com at The University of Iowa Libraries on June 19, 2015
 Effects of percutaneous edge-to-edge double-orifice repair under cardiac cycle in comparison with suture repair 967

Table 3 Model 1: MitraClipÒ device, leaflets principal extent than that generated by the clip. The maxi-
peak stresses (all values in MPa) mum stress occurs at the point indicated as AS in
the picture.
Diastole Systole Stress Equiv. alt.
range Von Mises
stress
3.2.2 Systolic phase
Point AC
sI: atrial side 0.19 20.05 0.24 0.17 Because, during systole, the leaflets are compressed
sII: atrial side 0.14 20.25 0.39 against each other [7], and the surgical suture is a
sI: ventricular side 20.04 0.12 0.16 0.07
sII: ventricular side 20.03 0.10 0.13 unilateral constraint that cannot oppose any inter-
nal action against compression, it is clear that the
Point BC
sI: atrial side 20.01 0.49 0.50 0.22 load transmitted by the suture can be considered to
sII: atrial side 20.03 0.23 0.26 be zero. The Von Mises stress contour map of the
sI: ventricular side 0.01 0.03 0.02 0.01
sII: ventricular side 0 0 0 leaflets under the effect of systolic pressure is
reported in Fig 10; the results are similar to those
obtained for the clip. In particular, the local effects
of the suture are not noticeable in this phase, and
substantially away from the clip. The local stress the maximum stress occurs in two symmetrical cen-
concentration due to the clip is therefore overcome tral parts of the anterior leaflet, in the point indi-
by the effect of the systolic pressure in the central cated as BS in the picture.
part of the leaflet body closer to the annulus. In Table 4, the same stress components as in
Table 3 are reported for points AS and BS.
Observing these results, it can be verified that the
3.2 Model 2: suture 4 mm width
low local effect exerted by the suture leads to small
stresses at point AS. As in the case of the clip, the
3.2.1 Diastolic phase
most highly stressed zone corresponds to point BS, in
The MVA and the transvalvular pressure gradient the central part of the anterior leaflet, with values very
(Dp) were determined using the same method as similar to that encountered in the model with the clip.
performed for the clip case. The predicted MVA
result is 2.03 cm2, and the transvalvular pressure
gradient is 7.30 mmHg. The total force transmitted 4 DISCUSSION
to the links at the end of the diastole is 0.25 N.
The Von Mises stress contour map of the leaflets The double-orifice technique is a relatively new sur-
under the effect of diastolic pressure is reported in gical repair technique, and the amount of published
Fig 9. The stress concentration effects near the data available is limited and often inconsistent
suture edges can again be noted, although at a lower between different authors, thus making it very

Fig. 9 Diastole with suture: leaflets Von Mises stress distribution (MPa)

Proc. IMechE Vol. 225 Part H: J. Engineering in Medicine

Downloaded from pih.sagepub.com at The University of Iowa Libraries on June 19, 2015
968 A Avanzini, G Donzella, and L Libretti

Fig. 10 Systole with suture: leaflets Von Mises stress distribution (MPa)

Table 4 Model 2: suture, leaflets principal peak stres- (about 9 mmHg) are in agreement with the peak
ses (all values in MPa) pressure gradient of 7.09 6 2.88 mmHg measured in
post-operative conditions after placement of the clip
Diastole Systole Stress Equiv. alt.
range Von Mises (see Table 5). Considering the MVA, it should be
stress emphasized that different methods are used to mea-
Point AS
sure it, which can result in significantly different
sI: atrial side 0.12 0.06 0.06 0.03 values, so it is more difficult to find consistent data
sII: atrial side 0.02 20.03 0.05 about valve pre- and post-operative conditions and
sI: ventricular side 20.01 0.15 0.16 0.07
sII: ventricular side 20.02 0.01 0.03 MVA values. Again, referring to reported clinical
Point BS
data [20], the predicted value of 1.83 cm2 is within
sI - atrial side 0 0.49 0.49 0.21 the range measured with the pressure-half-time
sI: atrial side 20.02 0.19 0.23 method (3.01 6 1.42) and the Gorlin formula
sII: atrial side 0.03 0.03 0 0.01
sI: ventricular side 0.01 20.01 0.02 (3.05 6 0.90), although closer to the lower limit, but
outside the lower limit of MVA measured by plani-
metry. Similar reasoning applies to the suture
model: the values are in agreement with the mean
difficult to compare experimental data, owing to the
MVA, assessed after surgical suture double-orifice
absence of the precise suture length and position,
repair, reported to be 2.3–2.7 cm2 (again depending
and of detailed information regarding the pre-
on the measurement method) in reference [21].
operative and post-operative conditions of the valve
Concerning the load transmitted by the leaflets to
itself. This study simulates the implantation of the
the link mechanism, it is practical to consider the
clip as an independent part applied to the mitral
diastolic and systolic phases separately.
valve, which is a new technique, and there are no
Under the effects of diastolic pressure, the load
direct terms of comparison regarding the force
was 0.39 N for the clip and 0.25 N for the suture.
transmitted to the MitraClipÒ device or the result-
This difference is due to the fact that the region of
ing stress on the leaflets.
the leaflets interacting with the clip arms is larger
Considering the above limitations, it is still worth
than the area near the leaflet edges on which the
comparing the FEM results with published literature
suture is applied, resulting in a larger area of leaflet
regarding this kind of surgical procedure, referring
to both clinical data and experimental in vivo and
in vitro testing. Since the clip and the suture have Table 5 MVA and diastolic pressure gradient compari-
been applied on the same valve model, it is also son for the two link typologies
possible to compare the behaviour of the valve in
MitraClipÒ Suture
the presence of different repair techniques.
The clip results can be compared with reported Diastolic MVA (cm2) 1.83 2.03
clinical measurements [20]. For the transvalvular Peak diastolic transvalvular 8.94 7.30
pressure gradient (mmHg)
pressure gradient the values predicted by the model

Proc. IMechE Vol. 225 Part H: J. Engineering in Medicine

Downloaded from pih.sagepub.com at The University of Iowa Libraries on June 19, 2015
 Effects of percutaneous edge-to-edge double-orifice repair under cardiac cycle in comparison with suture repair
constraint. It is interesting, however, to note that
the peak opening loads are of the same order of
magnitude as reported by Nielsen et al. [11] of
0.46 N in their in vivo ovine model. The Nielsen
model used direct measurement of the load on the
leaflet via a miniature force transducer, and the
measured peak load of 0.46 N occurred after indu-
cing ischemia. Similar values of peak suture tension
(0.30 6 0.18 N) are reported in reference [12]. During
systole, a differential compression force of 0.43 N is
transmitted to the clip, owing to the effective area of
its arms exposed to the ventricular side pressure,
which tends to close it. This result is valid under the
hypothesis that blood cannot penetrate between the
clip arms and the side of the ventricular leaflets side
– that is, that a perfect sealing is obtained between
these two elements. Nevertheless, there is no open-
ing load transmitted to the clip due to systolic pres-
sure application. No load at all is transmitted to the
suture, as it represents a unilateral constraint that
cannot bear compression loads. These results agree
with those reported by Umana et al. [22], who com-
ment: ‘During the high pressure of systole, the leaf-
lets are naturally pushed together with the suture
serving mainly to limit the excursion of either leaflet
into the atrium.’
In Table 6, the load peak values transmitted to
the link, representing the maximum and minimum
forces during the cardiac cycle, are shown.
A symmetrical push–pull alternate load cycle with
a maximum value of 0.39 N acts on the clip, while a
pulsating tensile load cycle with a maximum value
of 0.25 N is applied to the suture. Examining the
problem from the viewpoint of the linking mechan-
ism, these loads constitute external forces due to the
leaflets, which add to the internal forces due to the
link application, including suture pretension and
intrinsic forces within the clip due to the clip design.
Considering the leaflet stresses, very similar maxi-
mum equivalent alternating Von Mises stress values
during the cardiac cycle resulted for both the clip
and the suture cases, equal to 0.22 MPa and
0.21 MPa, respectively (see Tables 3 and 4). These
stresses are located in the central part of the ante-
rior leaflet on its atrial side, in two symmetrical cen-
tral zones of it. The reason for this similarity lies
Table 6 Link loads peaks (N) due to diastolic and sys-
tolic pressure application
MitraClipÒ
Diastole 0.39
Systole 20.43
Downloaded from pih.sagepub.com at The University of Iowa Libraries on June 19, 2015
969
in the fact that the maximum stresses are located
reasonably far away from the link, and therefore
are not influenced in a significant way by the link
typology itself.
In conclusion, the comparison of the clip and
standard suture techniques suggests that their func-
tional and structural effects on the mitral valve are
similar, so that in this sense they are comparable
from a clinical point of view.
4.1 Study limitations
Any model necessarily embodies simplifications and
limitations. For the present study a first set of lim-
itations concerns assumptions about the geometry
of the mitral leaflets. In particular, in the commis-
sure regions, the irregular shape of the leaflets,
tissue folding and redundancy, and the transition
from leaflet to chordae tissue were not considered.
As a consequence, the stress distribution in this area
and the shape of the orifice of the valve opening
may in reality be slightly different.
It has also been shown that annulus dilatation
reduces leaflet mobility in a sutured valve, thus
increasing the leaflet stresses slightly, whereas the
MVA and transvalvular pressure are not significantly
affected by this change [9]. The same behaviour is
to be expected with the clip, but more analyses are
necessary to ascertain it, and also to study more
severe pure annulus dilation compared with that
assumed in the present work.
Another fundamental aspect is represented by the
clip representation. In the present study the clip
was modelled as an independent part whose geome-
try reproduces the actual shape of the device.
However, in order to investigate the clip–leaflets
interaction more thoroughly, a solid (3D) element
model instead of present shell element model would
be more appropriate; this could also help in investi-
gating details of the local stress distribution around
the clip arms. As the main interest of the study con-
cerned the cyclic load transmitted to the link due to
the diastolic and systolic pressure application, some
simplifications have been introduced in the clip
model. In particular, at this stage of the research,
the mechanics of the clip itself have not been mod-
elled, so that internal forces, intrinsic to the clip
design itself, have been not taken into account.
Concerning the leaflet material, the choice of a clas-
sical hyperelastic constitutive law under the hypoth-
esis of isotropy has been discussed elsewhere [9,
Suture 17]; however, it has to be emphasized that the mate-
rial parameters were assumed to be uniform across
0.25 the valve leaflets, and the different mechanical
0
behaviours of the anterior and posterior leaflet
Proc. IMechE Vol. 225 Part H: J. Engineering in Medicine
970 A Avanzini, G Donzella, and L Libretti
tissue were treated as negligible. In reality, in addi-
tion to the anisotropy related to variation in the
principal fibre angles, the regional variation of fea-
tures such as collagen density or fibre tortuosity
may have significant effects on the leaflet material
behaviour, which have not yet been investigated for
the present model.
Leaflets soft tissues also exhibit a viscoelastic
behaviour, but this was not considered in this work,
owing to the lack of reliable stress relaxation data
for human heart valve tissues.
Finally, in this study the focus was on determin-
ing the effects of the clip on both the functional and
structural behaviour of the mitral valve. A key point
for the proposed procedure is the use of a Bernoulli-
type equation to couple the applied pressure, flow
rate and resulting valve opening. Although the equa-
tion coefficients for the specific case of the mitral
valve were derived by numerical fluid-dynamic
simulations, these are in turn based on simplified
assumptions on the three-dimensional behaviour of
the blood flow between the left atrium and ventricle,
which in reality is strongly dependent on the atrio-
ventricular geometry and kinematics.
5 CONCLUSIONS
In the present work, a finite element model is pre-
sented for studying the effects of the percutaneous
double-orifice repair technique of the mitral valve
using the MitraClipÒ device. The model, which is
based on previous work studying the structural
effects of the open-heart suture-based technique,
allows determination of the MVA and transvalvular
pressure gradient, as well as the structural effects of
clip application. In particular, the load transmitted
to the link created with the repair and the stress
state in the leaflets were evaluated, considering both
the diastolic and systolic phases of the cardiac cycle.
Details of the complex interaction between the
clip and the leaflet areas were also taken into account
by modelling the clip as an independent part with
dimensions corresponding to its actual size.
For comparison purposes, the same valve model
was also used to simulate the effects of a suture
having a width on the leaflet edge similar to that of
the clip.
Considering the MVA and the pressure gradient
across the valve, the results for the MitraClipÒ and
the standard suture are quite similar.
The leaflet stresses resulting from pressure varia-
tions during the cardiac cycle, which continuously
affect the mitral valve during working, are also very
similar for the two repair techniques, as the most
Proc. IMechE Vol. 225 Part H: J. Engineering in Medicine
Downloaded from pih.sagepub.com at The University of Iowa Libraries on June 19, 2015
highly stressed zones are located far away from the
link, in a position that is not affected by the link
mechanism.
Some differences were found regarding the load
transmitted to the link. In particular, in diastole, the
results showed that the loads are of the same order
of magnitude, but about 50 per cent higher for the
clip. A closing load is applied to the clip in systole,
with an absolute peak value comparable to diastole,
whereas no load is applied to the suture, owing to
its physical configuration.
As a consequence, under the effects of cyclic
pressure variation during the cardiac cycle, the clip
is subjected to an alternating push–pull external
load, whereas the suture is subjected to a pulsating
tensile external load. This consideration neglects the
effects of the load induced by the link application
phase, including internal forces intrinsic to the clip
design itself.
FUNDING
This work was supported by Evalve, Inc. (Abbott
Vascular, Santa Clara, CA).
Ó Authors 2011
REFERENCES
1 Maisano, F., Torracca, L., Oppizzi, M., Stefano, P.
L., D’Addario, G., La Canna, G., Zogno, M., and
Alfieri, O. The edge-to-edge technique: a simplified
method to correct mitral insufficiency. Eur. J. Car-
diothorac. Surg., 1998, 13(3), 240–246.
2 Alfieri, O., Maisano, F., De Bonis, M., Stefano, P.
L., Torracca, L., Oppizzi, M., and La Canna, G. The
double-orifice technique in mitral valve repair: a
simple solution for complex problems. Eur. J. Car-
diothorac. Surg., 2001, 122(4), 674–681.
3 Kuduvalli, M., Ghotkar, S. V., Grayson, A. D., and
Fabri, B. M. Edge-to-edge technique for mitral
valve repair: medium term results with echocardio-
graphic follow-up. Ann. Thorac. Surg., 2006, 82(4),
1356–1361.
4 Fucci, C., Cicco, G. D., Chiari, E., Nardi, M.,
Faggiano, P., Procopio, R., Coletti, G.,
Rambaldini, M., and Lorusso, R. Edge-to-edge
mitral valve repair for isolated prolapse of the ante-
rior leaflet caused by degenerative disease. J. Cardi-
ovasc. Med., 2007, 8(5), 354–358.
5 Tops, L. F., Kapadia, S. R., Tuzcu, E. M.,
Vahanian, A., Alfieri, O., Webb, J. G., and Bax, J. J.
Percutaneous valve procedures: an update. Curr.
Probl. Cardiol., 2008, 33(8), 417–457.
6 Davidson, M. J. and Cohn, L. H. Surgeons’ per-
spective on percutaneous valve repair. Coron.
Artery Dis., 2009, 20(3), 192–198.
 Effects of percutaneous edge-to-edge double-orifice repair under cardiac cycle in comparison with suture repair
7 Kar, S., Feldman, T., and St Goar, F. Mitral valve
repair using the MitraClip: from concept to reality.
Cardiac Interventions Today, 2008, May/June, 39–45.
8 Feldman, T., Kar, S., Rinaldi, M., Fail, P.,
Hermiller, J., Smalling, R., Whitlow, P. L.,
Gray, W., Low, R., Herrmann, H. C., Lim, S.,
Foster, E., and Glower, D. Percutaneous mitral
valve repair with the MitraClip system: safety and
midterm durability in the initial EVEREST (Endo-
vascular Valve Edge-to-Edge REpair Study) cohort.
J. Am. Coll. Cardiol., 2009, 54(8), 686–694.
9 Dal Pan, F., Donzella, G., Fucci, C., and
Schreiber, M. Structural effects of an innovative
surgical technique to repair heart valve defects. J.
Biomech., 2005, 38(12), 2460–2471.
10 Votta, E., Maisano, F., Soncini, M., Redaelli, A.,
Montevecchi, F. M., and Alfieri, O. 3-D computa-
tional analysis of the stress distribution on the leaf-
lets after edge-to-edge repair of mitral regurgitation.
J. Heart Valve Dis., 2002, 11(6), 810–822.
11 Nielsen, S. L., Timek, T. A., Lai, D. T.,
Daughters, G. T., Liang, D., Hasenkam, J. M.,
Ingels, N. B., and Miller, D. C. Edge-to-edge mitral
repair: tension on the approximating suture and
leaflet deformation during acute ischemic mitral
regurgitation in the ovine heart. Circulation, 2001,
104(12 Suppl. 1), I-29–I-35.
12 Timek, T. A., Nielsen, S. L., Lai, D. T., Tibayan, F.,
Liang, D., Daughters, G. T., Beineke, P.,
Hastie, T., Ingels, N. B. Jr, and Miller, D. C. Mitral
annular size predicts Alfieri stitch tension in mitral
edge-to-edge repair. J. Heart Valve Dis., 2004, 13
(2), 165–173.
13 Jimenez, J. H., Forbess, J., Croft, L. R., Small, L.,
He, Z., and Yoganathan, A. P. Effects of annular
size, transmitral pressure, and mitral flow rate on
the edge-to-edge repair: an in vitro study. Ann.
Thorac. Surg., 2006, 82(4), 1362–1368.
14 Kunzelman, K. S., Reimunk, M. S., and
Cochran, R. P. Flexible versus rigid ring annulo-
plasty for mitral valve annular dilation: a finite ele-
ment model. J. Heart Dis., 1998, 7, 108–116.
15 Kunzelman, K. S., Cochran, R. P., Verrier, E. D.,
and Eberhardt, R. C. Anatomic basis for mitral valve
modeling. J. Heart Valve Dis., 1994, 3, 491–496.
16 Maisano, F., Redaelli, A., Soncini, M., Votta, E.,
Arcobasso, L, and Alfieri, O. An annular prosthesis
for the treatment of functional mitral regurgitation:
finite element model analysis of a dog bone-
shaped ring prosthesis. Ann. Thorac. Surg., 2005,
79(4), 1268–1275.
17 Avanzini, A. A computational procedure for predic-
tion of structural effects of edge-to-edge repair on
mitral valve. J. Biomech. Eng., 2008, 130(3), 031015-
1:10.
Downloaded from pih.sagepub.com at The University of Iowa Libraries on June 19, 2015
971
18 Redaelli, A., Guadagni, G., Fumero, R.,
Maisano, F., and Alfieri, O. A computational study
of the hemodynamics after ‘edge-to-edge’ mitral
valve repair. J. Biomech. Eng., 2001, 123(6), 565–570.
19 Muntinga, H. J., van den Berg, F., Knol, H. R.,
Niemeyer, M. G., Blanksma, P. K., Louwes, H.,
and van der Wall, E. E. Normal values and repro-
ducibility of left ventricular filling parameters by
radionuclide angiography. International Journal of
Cardiac Imaging, 1997, 13(2), 165–171.
20 Herrmann, H. C., Rohatgi, S., Wasserman, H. S.,
Block, P., Gray, W., Hamilton, A., Zunamon, A.,
Homma, S., Di Tullio, M. R., Kraybill, K.,
Merlino, J., Martin, R., Rodriguez, L., Stewart, W.
J., Whitlow, P., Wiegers, S. E., Silvestry, F. E.,
Foster, E., and Feldman, T. Mitral valve hemody-
namic effects of percutaneous edge-to-edge repair
with the MitraClip TM device for mitral regurgita-
tion. Catheter. Cardiovasc. Interv., 2006, 68(6), 821–
828.
21 Frapier, J. M., Sportouch, C., Rauzy, V.,
Rouviere, P., Cade, S., Demaria, R. G., Davy, J. M.,
and Albat, B. Mitral valve repair by Alfieri’s tech-
nique does not limit exercise tolerance more than
Carpentier’s correction. Eur. J. Cardiothorac. Surg.,
2006, 29(6), 1020–1025.
22 Umana, J. P., Salehizadeh, B., DeRose, J. J. Jr.,
Nahar, T., Lotvin, A., Homma, S., and Oz, M. C.
‘Bow tie’ mitral valve repair: an adjuvant technique
for ischemic mitral regurgitation. Ann. Thorac.
Surg., 1998, 66(5), 1640–1646.
APPENDIX 1
Notation
A valve commissure-to-commissure
diameter
B valve anterior-posterior diameter
material parameters for hyperelastic
Ci0
constitutive law
Di0 material parameters for hyperelastic
constitutive law
I1 first deviatoric strain invariant
Jel elastic volume ratio
MVA mitral valve area
N order of polynomial strain-energy
function
Dp transvalvular pressure gradient
Q blood flow rate across valve
U strain-energy potential
sia alternate principal stress components
saMises equivalent alternating Von Mises stress
Proc. IMechE Vol. 225 Part H: J. Engineering in Medicine