Off-pump transapical mitral valve replacement: evaluation
after one month†
Kenji Iinoa,†, Jessica Boldta,‡, Lucian Lozonschib, Anja Metznera, Jan Schoettlera, Rainer Petzinaa,
Jochen Cremeraand Georg Luttera,*
aDepartment of Cardiovascular Surgery, University Hospital Schleswig-Holstein, Campus Kiel, Kiel, Germany
bDepartment of Cardiothoracic Surgery, University of Wisconsin, School of Medicine and Public Health, Madison, WI, USA
* Corresponding author. Department of Cardiovascular Surgery, University Hospital Schleswig-Holstein, Campus Kiel, Arnold-Heller-Str. 7, 24105 Kiel, Germany.
Tel: +49-431-5974625; fax: +49-431-5971252; e-mail: email@example.com (G. Lutter).
Received 8 March 2011; received in revised form 19 May 2011; accepted 30 May 2011
OBJECTIVES: The present study investigates outcomes one month after implanting pigs with a novel mitral valved stent.
METHODS: A novel nitinol stent custom designed for this study included a bovine pericardial valve. Six pigs received a valved stent
into the mitral position by means of the transapical implantation technique. Afterwards, haemodynamic stability and valve function
were assessed, immediately after implantation (n= 6), 4 weeks (n=4) and 8 weeks (n=1) thereafter using transoesophageal echocardi-
ography (TEE), ventriculography and cardiac computed tomography (CT). Four of 6 surviving pigs were sacrificed at 4 weeks after
implantation and one at 8 weeks thereafter.
RESULTS: Optimal deployment and positioning were obtained in all but one animal. This animal died of unrecognized imperfect
valved stent positioning after 4 days. The average mean gradient across the new valves and the left ventricular outflow tract (LVOT)
were small. Mild regurgitation developed after valved stent deployment in one of six animals just after 1 h, and in none thereafter. All
animals exhibited normal haemodynamics after mitral valved stent implantation, and stability was maintained throughout the monitor-
ing period. Migration, embolization and paravalvular leakage were not evident in the remaining animals after 4 and 8 weeks. Gross
evaluation revealed that 50–70% of the atrial element was covered by tissue growth at 4 weeks/8 weeks.
CONCLUSIONS: This study demonstrates adequate deployment and anchorage of a unique, repositionable mitral valved stent. A good
valve function was revealed in animals observed for 4 weeks and in one pig after 8 weeks.
Keywords: Mitral valve • Stent • Transventricule • Transapical • Minimally invasive • Valve surgery
Mitral valve disease is the second most frequent native valve
disease, behind aortic valve stenosis, in Europe. Transcatheter
aortic and pulmonary valve replacement procedures have
already been performed in selected patients. This has prompted
investigations into the feasibility of transcatheter implantation
with atrioventricular valved stents. Indeed, von Segesser et al.
described transcatheter mitral valved stent implantation , and
Boudjemline et al.  reported that placement of a valved stent
into the tricuspid position is feasible. However, some difficulties
remain to be resolved to develop and secure a valved stent in
the mitral position. First, adequate intracardiac imaging for stent
deployment with echocardiography or fluoroscopy is mandatory.
Second, a valved stent must avoid LVOT obstruction or radial
force by anchoring into the mitral valve annulus. Thirdly, a deliv-
ery system must overcome obstacles posed by the mitral valve
apparatus, which might impede perfect deployment. We
recently published our initial experience with acute transapical
off-pump mitral valved stent implantation with valved stents
comprising a flat star-like disk and a tubular ventricular piece
carrying different types of valves [3, 4]. These studies demon-
strated the feasibility of transapical off-pump mitral valved stent
implantation without LVOT obstruction. We also demonstrated
that this type of stent preserved adequate function without signs
of stent migration or embolization, systolic anterior movement
(SAM) or LVOT obstruction during one week of follow-up .
However, paravalvular leakage that emerged during the week of
follow-up and which was not observed at the time of stent de-
ployment remained an issue.
The novelty of this study is the midterm follow-up and de-
scription of our subsequent experience with a novel stent which
was evaluated at one month after off-pump transapical mitral
valve implantation in pigs.
†Presented at the 25th Annual Meeting of the European Association for
Cardio-Thoracic Surgery, Lisbon, Portugal, 1–5 October, 2011.
‡Contributed equally to this work.
© The Author 2012. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
European Journal of Cardio-Thoracic Surgery 41 (2012) 512–517
MATERIALS AND METHODS
We constructed an electropolished, radially self-expanding valved
stent from 0.3-mm nitinol wire. The stent comprised a flat ring
(atrial fixation system) connected at a 45° angle to a tubular ven-
tricular portion that accommodates a manufactured, trileaflet
bovine pericardial valve (diameter 28 mm) and a ventricular fix-
ation system. The ventricular fixation system consists of neo-
chordae between the base of the stent and the apex. The atrial
and ventricular elements had spontaneous diameters (at room
temperature) of 58 and 28 mm, respectively. A waterproof mem-
brane was sutured onto the atrial ring and over the ventricular
component to guarantee sealing, minimize paravalvular leakage
and allow easier repositioning. The diameter and length of the
crimped valved stent were 10.5 and 36 mm, respectively (Fig. 1).
Valved stent delivery system
The valved stent was inserted with a self-constructed, flexible,
catheter-based delivery system through the apex of the heart.
The external diameter and total length of the delivery catheter
were 12 mm and 32.5 cm, respectively. The delivery system com-
prised a 5-cm-long, customized front-loading sheath. The valved
stent was folded and housed using a crimping device in the
proximal portion of the catheter delivery system [4, 5]. An intern-
al pusher was attached to allow for smooth transition between
the tip and the distal part of the proximal capsule and to add-
itionally stabilize the entire sheath during its course. Balloon
dilation was not required to deliver the nitinol stent as it slowly
emerged from the sheath into the left atrium and ventricle.
All animals received humane care, as approved by the Center for
Experimental Animal Research at the University of Kiel (Kiel,
Germany) and in compliance with the Guide for the Care and
Use of Laboratory Animal Resources of the National Research
Council published by the National Academy Press (revised
1996). Six pigs weighing between 44.5 and 55 kg were implanted
with transapical off-pump valved stents.
Data were collected to evaluate haemodynamic stability for 60
min after implantation, 4 and 8 weeks and to assess the function
of the valved stents. Function was assessed by TEE at the time of
implantation, and at 1 h, 4 and 8 weeks thereafter. After prede-
termined time intervals, the four pigs were sacrificed and the
heart was macroscopically evaluated. Contrast ventriculography
proceeded only after 1 h in all pigs after complete deployment
of the valved stent.
The position of the stent in the four pigs at 4 weeks and in
one pig at 8 weeks after deployment was evaluated by CT using
an MDCT Somatom multi-slice CT scanner (Siemens, Erlangen,
Figure 1: (a) Bottom view of the new unfolded mitral valved stent, (b) operative setting after ministernotomy with this valved stent (bottom view) and apex of
heart with purse-string sutures is exposed.
Figure 2: (a) Transoesophageal echocardiographic view of mitral valved stent
along left ventricular long axis view at 90° angle. Pericardial valved stent is
accurately positioned. (b) Pulse wave Doppler flow across new valve shows
K. Iino et al. / European Journal of Cardio-Thoracic Surgery
Germany) at 64 slices per min, increments of 0.6 and the
CARDIAC-protocol with retrospective triggering (test bolus, 5 ml
Imeron 4 in LA; bolus, 1 ml Imeron 400/kg; reconstruction, MIP
0–100 in 5% steps).
The maximum diameter of the mitral valve annulus was mea-
sured (mid-oesophageal four-chamber, two-chamber and long-
axis view). The delivery system was positioned and the valved
stent was deployed across the mitral valve by TEE. Mitral regurgi-
tation and possible LVOT obstruction caused by the placed stent
were assessed by colour flow Doppler and pulse wave Doppler
sonography at the same time of TEE investigation.
The pericardium was opened after a lower mini-sternotomy was
performed through a 5- to 6-cm skin incision to expose the
apex of the heart. Two rows of 3-0 polypropylene pledgeted felt
purse-string sutures were then positioned. A bolus of 4000 units
of heparin was administered intravenously. The valved stents
were unloaded from the introducer device by a double-stage
procedure under TEE guidance (Fig. 2).
The left ventricular long axis was brought into view by TEE at
a 0° angle and the atrial component of the stent was partially
deployed firstly by advancing the pusher towards the tip of the
delivery system. During this phase of the procedure, the position
of the delivery system could be adjusted so that the atrial
component of the partially deployed valved stent was exactly
above the mitral valve annulus. A 140° view was also used for
subsequent valved stent positioning. The remaining part of the
stent was deployed by pulling back the delivery sheath while
holding the pusher in place. A full TEE examination of the valved
stent was planned immediately after deployment and after 1 h
and 4 weeks thereafter in all pigs, in one pig after 8 weeks.
Electrocardiography (ECG) was performed, and heart rate and
blood pressure were recorded continuously for 6 h post-
implantation. Contrast left ventriculography was performed
through the left ventricular apex after 1 h. In this study the
animals received no anticoagulation after the procedure.
All but one animal survived after correct deployment of the
mitral valved stent. One pig died due to unrecognized valved
stent mal-positioning that occurred 4 days after implantation.
Therefore, the results presented here are from the remaining
five pigs after the various observation periods. All animals exhib-
ited normal haemodynamics immediately after mitral valved
stent implantation and maintained stability for 6 h of monitoring.
Atrial and ventricular ectopic beats occurred during preparation
for apical access and valve deployment in all animals. No
arrhythmias were sustained or haemodynamically relevant in the
five remaining pigs.
Figure 3: (a) Three-chamber view of CT investigation 2 months after procedure with valved stent in correct position and no LVOT obstruction. (b) Top view of
atrio-ventricular valved stent with closed tricuspid valve in place. (c) Two-chamber view with closed pericardial valve at systole and left atrium beneath. (d) Lateral
view of frame of mitral valved stent.
K. Iino et al. / European Journal of Cardio-Thoracic Surgery
The mean surgical preparation time for apical access was 41
min (range, 35–46 min). All animals were implanted with a
bovine pericardial valved stent without any technical failures.
The mean mitral valve annular diameter was 29.2 (range, 27–31)
mm. The mean deployment time of the valved stent was 192
(range 153–243) s. Bleeding caused by insertion of the catheter
delivery system into the apical heart was generally controlled
easily. Blood loss was 81.3 (range, 35–110) ml on average and
occurred mostly during stent deployment. Two of the five valved
stents had to be repositioned twice after complete retrievement.
All the valved stents were accurately deployed after repositioning
and no stent migration or embolization occurred (Fig. 3). In add-
ition, LVOT obstruction and SAM did not occur in any of the
animals. The average mean transvalvular gradient across the
mitral valve after deployment was 1.8±0.2 and after 4 weeks
was 4.2± 0.3. The mean gradient across the LVOT immediately
after deployment (n=6) was 1.5 ±0.7 and after 4 weeks was (n=
4) 2.1±0.5. See also Tables 1 and 2. Contrast left ventriculog-
raphy revealed no mitral regurgitation in five of the six pigs and
mild paravalvular regurgitation in one pig after 1 h. Mitral insuffi-
ciency occurred in only one animal (Table 3). In this study the
animals did not receive any anticoagulation after the procedure.
No indication of TEE changes was evident throughout a 6-h ob-
servation period. At 4 weeks (n=4) and 8 weeks (n=1) after
evaluation, TEE demonstrated the absence of mitral paravalvular
leakage in all analysed animals (Table 3). Both CT and TEE in
four pigs at 4 weeks after implantation showed good positioning
of the stent, no LVOT obstruction and no mitral regurgitation
(Table 3). However, fracture of part of the atrial element of the
stent at the septal site was suspected in two of five pigs accord-
ing to CT at 4 weeks after implantation. Of them, one pig each
was sacrificed at 1 and 2 months after implantation. Gross as-
sessment of the five pigs after sacrifice demonstrated correct
stent positioning, no thrombus formation in all pigs and some
fracture of the atrial elements of the stent located at the septal
side of the left atrium in two of five pigs. The atrial elements
were covered by tissue growth and were confluent with 50% of
the surface area of the native atrial wall and the margins of the
atrial elements at 1 month and with 70% at 2 months (Fig. 4).
Animal 1 h after Mitral valve gradient (max)
MVG: maximal transvalvular gradient (mmHg).
Animal 1 h afterLVOT gradient (max)
LVOT: gradient across the left ventricular outflow tract (mmHg).
Because the Doppler left ventricular outflow tract interrogation
cannot be performed from a transgastric position in the swine model,
we used the angle correct feature of the Doppler packet to
compensate for the axis variation in the mid-oesophageal long-axis
plane for the estimation of LVOT gradients.
AnimalBefore Mitral regurgitation
1 h afterFour-week/8-week follow-up
LVOT: gradient across the LVOT (mmHg).
aNot applicable due to death after 4 days.
Figure 4: Nitinol mitral valved stent at 8 weeks after implantation. Top view:
left atrium at necropsy with signs of considerable circular tissue in-growth
K. Iino et al. / European Journal of Cardio-Thoracic Surgery
Although conventional open-heart valve surgery is the gold
standard for definitive surgical treatment of diseased heart
valves, some patients might not benefit from the procedure
because of a potentially high operative risk profile and strong
contraindications due to advanced age or significant comorbid-
ities. Transcatheter valve surgery allowing for off-pump beating
heart valved stent implantation has been developed over the last
decade as a minimally invasive alternative to conventional heart
valve surgery using cardiopulmonary bypass. These procedures
have already been conducted in selected patients with aortic
and pulmonary disease and has gained increasing acceptance
for high-risk patients.
Current percutaneous mitral valve repair techniques have the
principal limitation of being unable to fully correct most mitral
valve pathologies .
Transcatheter mitral valved stent implantation has the poten-
tial to achieve post-procedural results similar to those of surgical
repair or replacement. The first transatrial mitral valve implant-
ation of a double-crowned valved stent was described by von
Segesser and colleagues in 2005 . A self-expanding valved
stent intended for the mitral valve was deployed in a porcine
model via a conventional left thoracotomy and left atrial incision.
The valved stent remained in mitral valve position for 30 min fol-
lowing stent deployment. They used porcine aortic and pulmon-
ary valves that were preserved in glutaraldehyde as we did in
our first series of transapical mitral valved stent implantations .
The focus of our initial study concentrated on technical feasi-
bility. We investigated whether the star-like valved stents can
maintain correct positioning and whether such stents can over-
come LVOT obstruction that generally arises due to the ventricu-
lar portion of the valved stent protruding into the left ventricle,
which pushes the anterior leaflet of the mitral valve towards the
LVOT, mimicking an SAM effect . In subsequent series, we
demonstrated that off-pump transapical implantation of tricuspid
bovine pericardial valved star-like stents is feasible with good
valvular function, stable haemodynamics and no signs of either
LVOT obstruction or SAM for 7 days [4, 5].
However, paravalvular leakage that was not evident at stent
deployment emerged during the first week of follow-up .
When considering the use of such devices in patients with mitral
valve disease, minimizing paravalvular leakage after stent im-
plantation into the mitral valve will be a key issue to overcome,
as it would directly deteriorate life expectancy and quality of life
. It is well known that the morphological correlation between
the mitral valve and the LVOT is slightly different in humans
compared with that in pigs. Therefore, how these valved stents
will react in the human anatomy cannot be foreseen.
Therefore, we designed a new stent with a uniformly broa-
dened atrial element directly connected to a ventricular tubular
stent. The atrial element fully covers the mitral annulus from the
left atrium, which surmounts the issue of paravalvular leakage.
We demonstrated the absence of paravalvular leakage in four
pigs at 1 month and in one pig at 2 months after valved stent im-
plantation with good stent positioning and no LVOT obstruction.
However, CT findings indicated stent fracture in two of five pigs at
4 weeks after implantation. Gross assessment after sacrifice con-
firmed fracture of the atrial element at the septal site in these two
pigs. These fractures can be avoided if the atrial elements of the
valved stent are better aligned to the saddle-shaped structure of
the mitral annulus on the septal side. Furthermore, the mean gra-
dient of the valved stent increased moderately after 4 weeks com-
pared with baseline values. This could be due to the very small
paravalvular leakages occurring after stent fractures. Normally,
stent fractures of intracardiac valved stents are less common:
Bonhoeffer and his group demonstrated a fracture frequency of
21.1% after percutaneous pulmonary valved stent implantation.
Stent fractures over time were: after 1 year 14.9%, after 2 years
25.5% and after 3 years 30.8% .
However, tissue had covered and filled the gap between the
margins of the atrial element and the native atrium at 50 and
70% of the surface of the atrial element at 1 and 2 months, re-
spectively. The margins of the atrial element that were confluent
with the native tissue included part of the stent fracture in both
pigs. Furthermore, the valved stents were not thrombogenic
even in the absence of anticoagulation therapy. We do not know
precisely when the atrial element became fractured. The ana-
tomical features of the atrium are such that the space at the
medial site is relatively smaller than that at the lateral site. This
could explain the fracture of the atrial element at the septal site.
Further modification of the stent design is required.
We speculate that tissue growth on the atrial element will play
an important and indispensable role in filling the gap between
the stent and native tissue and to create a new smaller annulus
instead of the dilated native annulus in the clinical setting. The
healing response to this valved stent remains unknown.
However, studies on biological responses to the other intracar-
diac devices such as atrial septal defect occluders in humans and
other animals have shown that these devices are completely
covered by endothelial cells after 3 months [9, 10]. The results of
these studies are encouraging and indicate that a similar healing
process is achievable with our valved stent. However, further
studies including histological examinations will be required.
The long-term durability and fixation stability of these pericardial
valved stents are unknown because of the relative short-term
nature of this study. Moreover, the transcatheter mitral valved
stent was implanted only in healthy pigs with normal valves.
Another limitation of this study is that we did not look for any
cerebral thrombo-embolism after valved stent implantation.
Neurological deficits of the animals were not detected.
This study showed that an implanted pig could survive for
2 months. Initial preliminary animal studies had high complica-
tion and mortality rates caused by stent dislodgement, migration,
difficulties with advancing the catheter delivery system and para-
valvular leakage. Progressive improvements in design and de-
ployment techniques offer the promise of further advances with
The new mitral valved stent can be reproducibly deployed to
achieve reliable stent stability, minimal gradients across the LVOT
and adequate valved stent function without paravalvular leakage
for at least 1 month. Long-term function of the new valved stent
remains to be established.
K. Iino et al. / European Journal of Cardio-Thoracic Surgery
Dr Lutter’s project on transapical mitral valve replacement is sup-
ported by the German Research Foundation, Bonn, Germany
(Grant LU 663/8-1).
Conflict of interest: none declared.
 Ma L, Tozzi P, Huber CH, Taub S, Gerelle G, von Segesser LK.
Double-crowned valved stents for off-pump mitral valve replacement.
Eur J Cardiothorac Surg 2005;28:194–8. ; discussion 198–9.
 Boudjemline Y, Agnoletti G, Bonnet D, Behr L, Borenstein N, Sidi D et al.
Steps toward the percutaneous replacement of atrioventricular valves; an
experimental study. J Am Coll Cardiol 2005;46:360–5.
 Lozonschi L, Quaden R, Edwards NM, Cremer J, Lutter G. Transapical
mitral valved stent implantation. Ann Thorac Surg 2008;86:745–8.
 Lutter G, Quaden R, Osaki S, Hu J, Renner J, Edwards NM et al.
Off-pump transapical mitral valve replacement. Eur J Cardiothorac Surg
 Lutter G, Quaden R, Iino K, Hagemann A, Renner J, Hümme T et al.
Mitral valved stent implantation. Eur J Cardiothorac Surg 2010;38:
 Christofferson RD, Kapadia SR, Rajagopal V, Tuzcu EM. Emerging trans-
catheter therapies for aortic and mitral disease. Heart 2009;95:148–55.
 Genoni M, Franzen D, Vogt P, Seifert B, Jenni R, Künzli A et al.
Paravalvular leakage after mitral valve replacement: improved long-term
survival with aggressive surgery? Eur J Cardiothorac Surg 2000;17:14–9.
 Nordmeyer J, Khambadkone S, Coats L, Schievano S, Lurz P, Parenzan G
et al. Risk stratification, systemic classification, and anticipatory manage-
ment strategies for stent fracture after percutaneous pulmonary valve
implantation. Circulation 2007;115:1392–7.
 Kuhn MA, Latson LA, Cheatham JP, McManus B, Anderson JM, Kilzer KL
et al. Biological response to Bard Clamshell septal occluders in the
canine heart. Circulation 1996;93:1459–63.
 Sigler M, Jux C. Biocompatibility of septal defect closure devices. Heart
APPENDIX: CONFERENCE DISCUSSION
Dr M. Palmen (Leiden, Netherlands): My first question concerns the gradients.
You demonstrated very low inflow gradients after 1 h, which approximately
doubled after 4 weeks. And if I remember correctly, in one animal the gradi-
ents increased even further. Do you see any problems with this in the future
and could you comment on the underlying mechanism?
The second question is regarding the LVOT gradients, again very low LVOT
gradients demonstrated. And, as you noted, the anatomy of the aorto-mitral
continuity is different in pigs than it is in humans. Do you expect a higher
gradient in the LVOT if you implant it in humans?
Finally, can you comment on the future application of this excellent new
technique? Do you think it will replace open mitral valve repair, which we
know to have excellent results when using Carpentier’s techniques.
Dr Lutter: Yes, we are concerned about the transvalvular gradient; it was
?4.5 mmHg after 1 and 2 months, and this is the kind of restriction we see in
the pulmonary arterial circulation. We are trying to reduce this by reducing
the ventricular relaxation phase and to improve ventricular relaxation and get
it to normal. We have made improvements in the material which are helping
to reduce this transvalvular gradient.
It is not, I think, due to the valve, but rather to undersizing. Therefore, we
have to oversize in this respect and keep improving the material so that it
does not restrict movement of the left atrium and ventricle. This will surely
be a challenge but we are getting much better results right now, acute results
in other pigs with different valve stents.
In respect of your second question concerning the LVOT gradient, I think
up to now we do not see major problems despite SAM. SAM might arise in
humans. And yes, you have to find out the best height for the valve stent
body so that you do not disturb flow on the LVOT. We are familiar with this
problem, and you also have to avoid any shifting of the mitral valve stent
body into the LVOT, so this point of concern is also well taken.
And the third question was regarding future steps with humans. First, in
man, you have to see how the mid- and long-term data turn out. You might
also think of a disease model so that you are aware of all the problems you
might encounter, such as enlarged left atrium or left ventricle. This might be
more helpful when implanting such a valve stent by reducing obstruction to
the LVOT, as occurs with small left atria and ventricles.
So yes, we are looking forward. And as to whether we can replace annulo-
plasty or mitral valve repair techniques, I do not think so at this stage. This
will be a procedure for high-risk patients with numerous co-morbidities, so
we should not be afraid of this. We as cardiac surgeons should stay with this
technique and keep it in our hands, shouldn’t we?
Dr F. Maisano (Milan, Italy): You always show us new stuff in this field; you
are really a pioneer. I have a question for you regarding the method of fix-
ation. You have this method of implanting neochordae or, let’s say, four
tendons to keep the valve in position.
implant this valve in a patient with a severely depressed and enlarged LV?
What happens if there is a reverse remodelling and the chords now become
too long? Do you think this could be a problem? Do you think you can
undersize the length to overcome this issue?
Dr Lutter: Very good question, due to the fact that we always worry about
any migration of the valve stent. And yes, you have to seriously consider and
face this problem at an early stage. What we do is concentrate on different
anchoring mechanisms: one is best alignment in the left atrium, and two is
having a good seat of your atrial elements in the left atrium and a mild radial
force, so little oversizing of the valve itself in the mitral annulus plane, and
third, you have the tethering, which means that sometimes you don’t need to
pull it much to fix this if you have it in the right position. And this has to be
always improved. And right now we have, as I said, much improved materials
and then you don’t need so much undersizing of the chords to fix your valve
stent. I would not be too concerned about what happens in animals as, even
in severe left ventricle disease in patients, the situation is more forgiving and
a device therefore easier to implant.
What is your thought once you
K. Iino et al. / European Journal of Cardio-Thoracic Surgery