Left Atrial Appendage Ligation And Exclusion Technology In The Incubator
Faisal F. Syed BSc (Hons.), MBChB, MRCP,1, Amit Noheria MBBS,1, Christopher V. DeSimone MD, PhD,1, Samuel J. Asirvatham MD, FACC, FHRS1,2
Division of Cardiovascular Diseases, Department of Internal Medicine, Mayo Clinic, Rochester, Minnesota.2Department of Pediatrics and Adolescent Medicine, Mayo Clinic, Rochester, Minnesota.
Stroke is the most feared complication of atrial fibrillation (AF). Targeting the left atrial appendage (LAA) mechanically is attractive as a means to simultaneously reduce stroke risk, the need for anticoagulation, and hemorrhagic complications in patients with non-valvular AF. The results of the PROTECT-AF and PREVAIL randomized clinical trials support this approach as a viable therapeutic alternative to warfarin in selected patients and add to accumulating evidence regarding the importance of the LAA in thromboembolism in AF. A number of devices for percutaneous LAA closure are under investigation or development. In this article, key design features of these ligation and exclusion technologies will be discussed, with a focus on aspects of LAA morphology, relational anatomy, thrombosis, and thromboembolism relevant for successful device development and deployment.
Key Words : Left Atrial Appendage, Ligation, Exclusion, Technology, Anatomy, Thromboembolism, Device.
Corresponding Address : Samuel J. Asirvatham, MDProfessor of Medicine and Pediatrics, Division of Cardiovascular Diseases200 First Street SW, Rochester, MN 55905.
Left atrial appendage (LAA) closure, by altering the balance of Virchow’s triad within the appendicular cavity,1 is an attractive strategy for stroke prevention in nonvalvular atrial fibrillation(AF).2-5 An appreciation of the pathophysiological influence of the LAA on stroke risk traditionally hinged on the observation that 15% of patients with nonvalvular AF have intracardiac thrombi, of which 90% are located within the LAA.6-9 Recent advances in cardiac imaging have allowed investigators to demonstrate that morphological complexity of the LAA significantly influences thromboembolic risk, supporting a structural approach to thromboprophylaxis.10-16 This principle was tested by the PROTECT-AF and PREVAIL randomized clinical trials, which demonstrated that LAA exclusion using the WATCHMAN percutaneous occlusion device (Boston Scientific, Natick, MA) was not clinically inferior to warfarin in preventing strokes, whilst reducing bleeding risk.3, 17 Till now, the only percutaneous LAA closure device available in the USA for ameliorating AF-related stroke risk has been the LARIAT appendage ligation system (Sentre-HEART, Redwood City, CA),18-22 which provides an attractive alternative approach for stroke prevention in patients with a high bleeding risk on systemic anticoagulation. A number of other technologies have received CE (Conformité Européene) mark approval for commercial use in Europe, of which the AMPLATZER cardiac plug (ACP, St Jude Medical, Saint Paul, MN)23-27 and the now discontinued PLAATO system (eV3, Sunnyvale, CA)28-30 have been the most widely implemented. Others are currently in the incubator, although reporting preclinical or early clinical results (Table 1).
In this article, aspects of LAA morphology, relational anatomy, thrombosis, and thromboembolism relevant for successful percutaneous LAA closure will be discussed initially, drawing from published observations on devices currently in clinical use. Subsequent focus will be on key design features of devices under clinical investigation or development.
Endpoints For Effective LAA Closure: Epicardial Ligation Versus Endocardial Exclusion
As a broad overview, devices have utilized either an endovascular exclusion-based approach, in which a foreign occlusive body is introduced via atrial transeptal puncture and deployed within the LAA, thereby excluding it from the main atrial chamber (WATCHMAN, ACP, and PLAATO), or an epicardial ligation-based approach where epicardial puncture permits navigating to and tying down a noose over the LAA neck (LARIAT).4, 31-36 Aside from specific variations in device design and application, which translate into important procedural and patient selection considerations, the resulting structural changes that follow are observably dissimilar between these approaches (Figure 1).37, 38 The endpoints for adequate
closure are therefore also likely to differ, and whether adequate closure
is best defined using anatomical, electrical, or functional criteria4, 31
will ultimately depend on recognizing the respective association with
clinical effectiveness. Data from surgical ligation suggests that both
residual leak from incomplete ligation and residual stump from too
distal a ligation point predispose to subsequent atrial thrombus.39-43
Late LAA leakage has been reported in 20% to 25% of patients at
3 months with LARIAT,21, 22, 44, 45 and a predisposition to subsequent
atrial thrombus formation is speculated46 though not proven. Late
thrombus formation at the site of closure has been reported in 5%
of patients between 17 to 104 days after LARIAT ligation22 and
both in the presence46 and absence of a residual leak.47 Peri-device
leaks following endovascular occlusion are also common, being
noted in 41%, 34% and 32% at 45 days, 6 months and 12 months
respectively in patients from PROTECT-AF after WATCHMAN
closure,48 although the currently published data suggest no associated
increase in rates of thromboembolism.48-51 The velocity of the leak,
the degree of residual exposed LAA anatomic complexity, and the
ability of the leak to accommodate a thrombus may be quite different
between patients with a ligation-related leak versus those with a
peri-device leak. Thrombosis mechanisms in device exclusions are
often related to the device itself (3.7% of patients in PROTECTAF
receiving WATCHMAN)32 and more likely to be influenced
by seating of the device within the LAA and subsequent device
endothelialization,52 with less complete endothelialization of the
ACP compared to WATCHMAN on account of the ACP’s larger
surfaced extra-appendicular disc and more prominent end-screw
hub in a comparative dog study with WATCHMAN (Figure 1).38
Ligation results in acute appendage ischemia leading to appendage atrophy and cavity obliteration5, 37, 53, 54 although the contribution
of this remodeling to preventing thrombus formation is currently
unknown. This is also the case with residual “beaks” where tissues are
approximated, residual diverticula or extra-appendicular pectinate
ridges.4, 31
Figure 1. Left and middle columns: canine specimens demonstrating comparative positioning of the WATCHMAN (left) and ACP (right), with histological sections (bottom) demonstrating cross sectional relationship to ostium and neo-endothelialization (inset). Note the intra-appendicular positioning of the WATCHMAN compared to the external disc of the ACP, and the incomplete endothelial coverage of the ACP mesh wires near the inferior edge of the disk and end-screw hub.38 Right column: Changes following LARIAT occlusion with suture (green) in the explanted heart of a patient who underwent cardiac transplantation 1 year 11 months after the procedure, with corresponding Mason trichrome staining (bottom) demonstrating extensive scarring (dark blue) within the appendage and suture site that extends into the left atrium .37 LSPV – left superior pulmonary vein. MV – mitral valve. LAA – left atrial appendage. LA left atrium. Endo – endocardial surface.

Table 1.
FDA-approved and CE mark | LARIAT |
---|
CE mark only | Amplatzer Cardiac Plug
Amplatzer Amulet
Transcatheter Patch
Watchman (FDA-approval applied for, awaiting decision)
Wavecrest |
Published preclinical or human studies | Aegis
Epitek (withdrawn)
LifeTech LAmbre (Phase 0 trial ongoing)
Ultrasept |
Others | Occlutech (Phase 1 trial registered, not yet recruiting) |
Key Considerations For Device Design
The LAA is a tubular projection arising from the free wall of the left
atrium, typically extending superiorly to project a variably curvilinear
course, bending noticeably in 75% individuals at 98 ± 20 degrees
after the initial 14 ± 4 mm, running adjacent and parallel to the left
superior pulmonary vein (LSPV), underneath the main pulmonary
artery, and draping down over the right ventricular outflow tract, left
main coronary artery bifurcation, left atrioventricular groove which
houses the left circumflex artery and great cardiac vein, and a portion
of the mitral annulus (Figure 2).55-61 The sinoatrial node artery can
be related when it arises directly from the left circumflex artery (30%
of individuals) or coursing from the left lateral atrial artery (8% of
individuals) rightward between the appendage and LSPV towards
the sinoatrial node (Figure 2).58 The left phrenic nerve runs along the
overlaying pericardium62 traversing the appendage variably from over
its tip to over the roof of the ostium.63
Figure 2. External anatomy of the LAA. Bottom: CT images with the appendage highlighted in pink, demonstrating variation in morphology and orientation.55, 63, 58 LAA – left atrial appendage; LSPV – left superior pulmonary vein; L-SANa - left sinoatrial node artery; LCx – left circumflex artery.

A transcutaneous epicardial approach to the LAA must therefore
first negotiate the anterior pericardial space, with free passage
superiorly to engage the appendage whilst avoiding the above
mentioned neurovascular structures. Such an approach to the LAA
may be restricted in individuals with pericardial adhesions from prior
open heart surgery, pericarditis, epicardial VT ablation, or uremia,
anatomical distortion such as with pectus excavatum, kyphoscoliosis
or severe obesity58, 64 or congenital abnormalities such as pericardial absence which may additionally be associated with appendage
herniation.65
Figure 3. Endocardial anatomy. Top right: Cross sectional histology of the left atrial appendage. The location of the ostium is denoted by the interrupted line.58 Top left: CT reconstruction of endocardial anatomy demonstrating the variation in position of the left lateral ridge (arrows, R) and left pulmonary veins.57 Bottom: Variation in size and shape of the left atrial appendage ostium.58 LAA – left atrial appendage; MV – mitral valve; LCx – left circumflex artery; LS – left superior pulmonary vein; LI – left inferior pulmonary vein

Detailed morphological LAA definition on an individual basis
is integral to success of epicardial ligation approaches as there is
marked inter-individual variation in appendage size, number of lobes,
crenelations, and pectinate musculature,56, 66, 67 with 20 to 70 percent of
individuals having a single lobe, 16 to 54 percent having two, and the
remainder having up to 4 lobes.68, 69 Occasionally a second posterior
lobe is sandwiched between the right and left ventricular outflow
tracts.55 Although the appendage tip points inferiorly in the majority,
usually pointing down in line with the course of the left anterior
descending artery, it can sometimes be directed posteriorly or into the transverse pericardial sinus (Figure 2).58 These considerations are
important as a noose-based ligature has to first engage and be slipped
past the appendage tip. Without an effective method to manipulate
or re-orientate the appendage tip, the LARIAT is unsuitable for
patients with a superiorly oriented LAA or a posteriorly rotated
heart.58 Ensuring the noose is large enough to capture the appendage
and any additional lobes is also key; the LARIAT’s noose has a
maximal diameter of 40 mm.18 Identifying the relationship of the
left main coronary artery and its bifurcation to the neck, and the
course of the phrenic nerve, are also important. Appendages may
in addition be closely adherent to the underlying left ventricular
wall, congenitally aneurysmal,70 inverted,71 juxtaposed,72 or absent
altogether.73
Figure 4. Appendage width and depth takes precedence to morphology for endocardial device closure.80

The above anatomical and pathophysiological constructs present
distinct, sequential challenges in executing a successful strategy for
percutaneous epicardial LAA ligation.4 First, in contrast to the ease
with which the appendage can be instrumented through transeptal
access, negotiating it accurately from within the pericardial space is
difficult even when guided by preprocedural CT and intraprocedural
TEE. Secondly, once identified, the freely mobile appendage requires
stabilization to allow for controlled ligation. LARIAT addresses
both these steps simultaneously by using an adjunctive transeptal
endocardial approach to first advance one magnet-tipped wire to
the appendage tip, and approximate another from the epicardium
onto it, thereby grasping the appendage whilst establishing a
supporting monorail over which to advance a looped suture.18
This unique approach does, however, introduce risks from both
transeptal and epicardial puncture, risk of appendage puncture from deep intubation of the appendage by the transeptal sheath, risk of
appendage laceration, and the procedural complexity of maintaining
correct orientation of two curved sheath-based platforms relative
to each other and the appendage wall.64, 74-76 Thirdly, although the
optimal ligation site along the LAA is not known, closure is targeted
at the ostium based on surgical data implicating persistent leaks and
residual stumps for recurrent thrombosis, yet the boundary between
the LAA and main atrial chamber is indistinct both anatomically
and electrically. To overcome this, both endocardial and epicardial
strategies attempt to close the appendage as proximally and snugly
as possible,33 which for LARIAT involves placing an endocardial
balloon-tipped catheter at the ostium, overriding it when tying
down the epicardial suture, and ensuring a tight seal using contrast
fluoroscopy and Doppler TEE.18, 64 However, the final tightening
is reserved for once the endocardial citing balloon is deflated and
withdrawn. Despite ensuring an adequate seal, which can be by
radiocontrast injection as well as Doppler color flow imaging, recent
clinical experience with the LARIAT has reported that leaks reoccur
and can be seen in 20% to 25% of patients within a few months,21, 77
cause or consequence is unknown at present.
Figure 5. WATCHMAN device, demonstrating active fixation barbs.105

The LAA can be identified internally by pectinate muscles, which impart the characteristic combed appearance of its endocardial
surface, although in a significant proportion of individuals these
can extend inferiorly to the vestibule of the mitral valve.56, 58, 63
The smooth-walled LAA ostium, whilst demarcated by the LLR
superiorly and posteriorly, has indistinct borders anteriorly and
inferiorly which therefore have to be approximated (Figure 3).56
Approaches for endocardial exclusion have relied upon TEEvisualization
of the left main coronary artery, circumflex artery or
mitral annulus to define the plane of the ostium in relation to the
LLR.33, 78-80 The shape of the LAA ostium is usually oval, rather than
round, though varies significantly between individuals,56, 66, 69 and the
orientation of the ostium relative to the plan of the mitral annulus is
oblique rather than vertical.56 The ostium can be at the same level as
the LSPV (60% to 65% of individuals), superior to it (25% to 30%) or
inferior (10% to 15%) (Figure 3).57, 69 The intervening LLR is formed
by an infolding of the lateral atrial wall, is narrow superiorly where
it is predominantly muscular, becomes up to 5 mm wide inferiorly
and houses the ligament of Marshall (remnant of the left sided
superior vena cava), autonomic nerves and a small atrial artery which sometimes is the sinoatrial nodal artery.56, 58, 62, 63 Distance from the
LAA ostium to the LSPV is 5-10 mm in 45% of individuals, and 10-
15 mm in 40%, can be up to 24 mm in the remainder, with distance
from LAA ostium to mitral annulus being similar.56
Figure 6. Aegis electrogram guided intrapericardial ligation approach. Left: Demonstration of intrapericardial navigation to the appendage tip using grabber and shaft electrodes, and preformed looped suture application. Right: From top to bottom on the tracing are shown surface ECG leads I, III, avF, jaw-to-jaw bipolar electrogram (EGM), jaw1-to-shaft EGM, jaw2-to-shaft EGM, and a shaft bipolar recording. Tracings on the left are pre LAA ligation, and on the right are post LAA ligation. Note the disappearance of the atrial signal from the appendage following ligation, and the predominant ventricular signal on the shaft bipolar electrogram.5

By engaging the ostium directly from within, the success of an
endocardial exclusion strategy is dependent on the interaction of
ostial size and shape with accurate device positioning and adequate
exclusion, the internal anatomy of the LAA relied upon by the device
for a safe and secure landing, and how well transeptal access orientates
engagement with the appendage ostium.33, 58 Devices design is
required to incorporate ways to minimize and compartmentalize
thrombosis such as by using resistant materials or minimizing device
profile and exposed structures, a delivery platform that facilitates
accurate and safe device positioning, a deployment mechanism
which can be adjusted for size and position to achieve an adequate
seal, a seating mechanism which prevents device embolization, and
a profile which does not interact significantly with related structures
including the left lateral ridge (LLR), mitral valve, pulmonary vessels
and coronary arteries.
Figure 7. The Amplatzer first generation ACP (left) comparted to second generation Amulet (right).84

The WATCHMAN is designed to be deployed 10 mm below
the LAA ostium such that its self-expanding nitinol cage fills
the appendicular cavity and its 160 μm thick covering made from
microporous polyethylene terephthalate entraps thrombi and
promotes endothelialization.31 The device is unfurled by gradual
pullback of the access sheath and delivery catheter while maintaining
device position using fluoroscopic and TEE guidance, with Doppler
flow to identify adequacy of the seal and the sheath to facilitate partial
recapture and adjustment.33 The ACP has a braided nitinol frame with
overlying polyester patch and is designed to cover the ostium with a
disc articulated to a distal lobe which anchors within the LAA.31
The ACP lobe is deployed first by partial unsheathing, followed by
the proximal disc by further unsheathing, whilst effectiveness of
occlusion is confirmed using distal radiocontrast injection through
the delivery system and/or Doppler flow78 and with partial or full
retraction into the sheath for repositioning.33 Currently available
devices are for ostial sizes of 17-31.9 mm for WATCHMAN and
12.6-28.5 mm for ACP.33
Figure 8. The Lifetech LAmbre device.97

In contrast to the epicardial approach, accurate and standardized
measurement of width of the ostium and depth of the landing zone
take precedence to other morphological considerations in ensuring
that the device is compatible with the appendage and correctly sized
(Figure 4).33, 80 The WATCHMAN requires the LAA length to be in
excess of the maximal ostial diameter and is therefore better suited
for long and narrow appendage profiles, whilst the ACP is better
suited for short and broad profiles as the anchoring lobe requires
the landing zone to be at least 10 mm wide.33, 80 Challenging
morphologies include appendages which taper significantly from
ostium to tip, where usual sizing of the ACP landing site may result in
an undersized disc at the ostium33 and “chicken wing” morphologies
which can have an excessively early and severe bend.81 Rarely, there
may be an ostial membrane manifesting with elevated gradient across
the ostium.82
Figure 9. The Occlutech LAA Occluder.98

Accurate appendage sizing is also important in ensuring a snug
fit and reducing risk of device embolization, and accordingly WATCHMAN devices are sized 10% to 20% larger and ACP 1.5-
3.4 mm larger than the maximal ostial diameter.33 However, relying
on radial expansion forces alone has been shown to be insufficient
in ensuring device stability: the very early AMPLATZER septal
occluders which relied on this strategy when deployed within the
LAA had high rates of device embolization.31, 78 AF is associated
with an increase in appendage size and reduction in the internal
trabecular structure due to pectinate muscle atrophy and endocardial
fibroelastosis83 whilst the ostium progressively increases in size
and adopts a more rounded shape with increasing AF burden.79 In
addition, with appendages which taper distally, radial forces may
paradoxically result in device expulsion as the pressures generated
deeper in the appendage will be greater than those at the ostium.33
Current devices use active fixation mechanisms which are engaged
using gentle application of negative traction upon deployment and
take advantage of the appendage’s trabeculated endocardium (Figure
5).33, 78 A strategy of oversizing devices serves also to reduce the
incidence of peri-device leaks,84 the significant incidence of which
is likely related to the variably oval shape of the ostium56, 66, 69 in
contrast to the uniform and rounded design of devices in current
clinical use.23, 48-51, 84, 85 However, care must be taken not to distend
aggressively, as the appendage is paper-thin in areas between the
pectinates and may perforate, and the ostium is critically located
immediately anterolateral to the left main coronary artery, superior
to the great cardiac vein and circumflex artery, and anterior to the
LSPV,56, 58, 63 any of which may become compressed. Reported
complications following endovascular exclusion procedures have also
included erosion into the overlying main pulmonary artery.86-88
Figure 10. The Transcatheter Patch.99

All current endocardial exclusion strategies utilize a sheath based
delivery platform with access across the interatrial septum.33, 80 Given
the ostium’s oblique orientation,56 a posteroinferior septal puncture
allows approaching the ostium at an optimal angle without excessive
sheath manipulation and torque, which increases risk of atrial
perforation.33, 58 Slight adjustment of the approach is required for
each case to accommodate for the inter-individual variation in the
angle adopted by the septum in its left anterior to right posterior
course.58 Puncture should be at the true anatomical septum which
is defined by the thin floor of the fossa ovalis, measuring 1-3 mm in
thickness, whilst the muscular rim is formed by invagination of the
atrial wall, though location and size vary between individuals and in
those with kyphoscoliosis and marked left ventricular hypertrophy63
Echocardiography-guided puncture is to be recommended given
the high prevalence of a septal ridge,57 pouch at the fossa,63 or other
structural abnormalities including atrial septal aneurysm, patent foramen ovale, atrial septal defect, septal flap, thickened interatrial
septum, or thrombus.89
Device Technology In The Incubator
The ideal AF stroke prevention technique should completely
remove any thromboembolic risk and substrate, confer minimal
clinical risk, be cost effective and applicable to all.4 Towards this goal,
the experience with WATCHMAN and ACP, and more recently
LARIAT, add further insight into determinants of success and
current shortcomings of device design and approach, even though no
study has to date directly compared one device to the other.90 With
development ongoing, opportunities arise for improving efficacy,
universal applicability, safety and simplicity.
Figure 11. The Cardia Ultrasept LAA Occluder.100

The Aegis system (Aegis Medical, Vancouver, Canada)5, 53, 91 is
a totally intrapericardial ligation approach which harnesses the
appendage as the most inferior site of atrial electrical activity obtained
from an anterior subxiphoid epicardial approach.4 A steerable
epicardial sheath, placed via standard subxiphoid puncture, supports
the introduction of an appendage grabber with embedded electrodes
within the jaws and further electrodes on the shaft proximally.
The grabber is electrically navigated onto the atrial appendage,
which it then captures and stabilizes, whilst ventricular signals on
the proximal shaft electrodes confirms an orientation towards the
appendage tip (Figure 6). A hollow suture preloaded with a support
wire to permit remote suture loop manipulation and fluoroscopic
visualization is advanced to the appendage base and looped around
the appendage, with a range of appendage sizes, shapes and lobes
enabled by the variable loop size. After loop closure, the wire is
removed, leaving only suture behind, which is remotely locked with
a clip to maintain closure. If initial closure is unsatisfactory, the
loop can be undone and repositioned, or additional loops placed over the first. Successful closure is confirmed within seconds by the
elimination of LAA electrical activity, accompanied by shortening
of the surface electrocardiographic P wave in dogs5 and followed by
the LAA becoming atretic.53 As compared to LARIAT, the major
advantage offered by Aegis is that transeptal access is not required and,
therefore, neither is anticoagulation. Similar to LARIAT, previous
cardiac surgery or adhesions from previous pericarditis are the major
limitations. Feasibility in humans has been demonstrated,91 with
approximately 50 patients having had the procedure to date.92
Figure 12. The WAVECREST Left Atrial Appendage Occlusion System. Courtesy of Coherex Medical, Salt Lake City, UT

A second generation ACP, the Amplatzer Amulet Left Atrial
Appendage Occluder (St. Jude Medical, Saint Paul, MN, USA)93-
95 received European CE Mark approval in 2013 although it is
currently voluntarily withdrawn from the USA market by St. Jude.
The design is similar to ACP, with a lobe-disc structure made from
nitinol mesh covered by polyester patches (Figure 7). In comparison
to the ACP, the Amulet is designed for superior seating whilst
requiring less oversizing, with a 2-3 mm longer lobe housing stiffer,
more evenly distributed and more numerous stabilizing barbs (from
six pairs in ACP to 10 pairs), and longer articulating waist between
the distal lobe and the proximal disc. Having larger available sizes
(31 and 34mm), it is better suited for closure of larger LAA. Aimed
at reducing device thrombosis, the screw facing the atrial chamber is
now flush to the device, and the larger disc designed to be seated flush
to the LLR and less prone to prolapsing into the ostium, which is
thought to predispose to thrombus formation with the current ACP
by creating a cul-de-sac with the LLR. To facilitate deployment,
it now comes pre-mounted on a modified pusher cable inside the
delivery system. Recently published non-randomized clinical
experience in 25 patients in Europe reported successful implantation
in 24, with complications of 1 device thrombosis; there were no leaks
>3 mm or other complications.96
The Epitek (Medford, NJ, USA) multilumen system utilizes a fiberoptic
endoscope, jaws to grasp the LAA visualized endoscopically, a
pre-tied suture, and shape-set nitinol wire. Testing in porcine and
canine models was performed from December 2006 to February 2008,
leading on to early human testing where difficulties were encountered
with access (achieved in 78%) and good device positioning (achieved
in 41%) with subsequent development halted.92
The Lifetech LAmbre device (Lifetech Scientific Corp., Shenzhen,
China)97 has some similarities to ACP, though in place of the lobe
there is a nitinol-based, fabric covered, self-expanding umbrella
which is introduced into the LAA (Figure 8). The umbrella is
secured within the LAA via 8 distal ball-tipped frames with side
facing hooks, and articulates via a waist to a disc which orientates
onto and seals the ostium. The device is deployed through a sheath,
and is retrievable and repositionable. After testing in a dog model,97
a feasibility and safety human study is underway (clinicaltrials.gov/NCT01920412) and CE Mark is expected in the near future once
adequate patient experience is obtained.
The Occlutech LAA Occluder (Occlutech International AB,
Helsingborg, Sweden) is based on a braided nitinol frame which is
introduced into the LAA (Figure 9).98 The contour tapers distally to
better distribute radial expansion forces and it is anchored distally
with closed loops designed to engage the trabeculated LAA whilst
avoiding perforation. A polymer covering seals against blood flow
and promotes endothelialization. Delivery is via an endocardial
sheath, and a ball-shaped connection hub allows the occluder to pivot
during delivery. Sizes of 17-39 mm are available. Phase 1 feasibility
and safety study has not started enrolling as yet (clinicaltrials.gov/
NCT02105584). No published data using this device is available
currently.
The Transcatheter Patch (Custom Medical Devices, Athens,
Greece)99 is also deployed endocardially within the LAA to occlude
it (Figure 10). The unique features are that it is frameless, being
made from bioabsorbable polyurethane foam and kept inflated by
radiocontrast to diameters of 15-25 mm. It is secured within the
LAA initially by polyethylene glycol glue, activated by an alkaline
solution followed by a prolonged (45 minute) inflation, and over the
subsequent 48 hours via fibrin formation. A 2-mm nylon loop is
sutured at the bottom of the patch, and a double nylon thread is
connected for retrieval purposes. The Transcatheter Patch has CEMark
approval for the use of occlusion of heart defects in general.
Feasibility in LAA occlusion was reported in 17 patients, although
in 3 the patch did not attach and in 1 it was placed beyond the LAA
ostium, whilst sheath thrombosis was seen in 1 patient.99 There were no strokes at 1 year follow-up.
The Cardia Ultrasept LAA Occluder (Cardia Inc, Eagan, MN)
is made from a nitinol frame with a distal cylindrical anchor which
is deployed endocardially within the appendage, secured therein
using 12 hooks strengthened onto platinum/iridium collars, thereby
providing support via a flexible articulation to a round sail made from
polyvinyl alcohol foam which orientates onto and covers the ostium
(Figure 11).100 The stranded design of its frame reportedly increases
fatigue resistance and allows fine tuning of the tension applied to the
sail and anchor, the long and flexible waist allows increases positional
versatility, whilst the sail is reported to be designed to minimize blood
flow disturbance within the LAA.100 Successful deployment in 5 dogs
has been reported with complete neointimal coverage on histology at
30 days.100 The device comes in five sizes for human use, based on
the diameter of the distal bulb:16, 20, 24, 28 and 32mm. No human data
is currently published.
The WAVECREST®101 Left Atrial Appendage Occlusion
System (Coherex Medical Inc., Salt Lake City, UT, USA) is nitinol
framed, Gore-Tex covered device (Figure 12) similar in principle to
WATCHMAN but with a number of design features to overcome
current WATCHMAN limitations.80, 98 It has an umbrella-shaped
frame designed for shallow deployment making it suitable for a wide
range of appendage sizes. This is coupled with less stringent sizing
criteria allowing for 3 sizes (22 mm, 27 mm, 32 mm), coverage for
ostial sizes of 18-30 mm, a completely retrievable and repositionable
sheath-based system, and a distal radiocontrast delivery system to
guide adequate positioning. Expanded polytetrafluoroethylene
(ePTFE), which has low thrombogenecity, covers the occluding
cap. Safety features include 20 anchoring nitinol microtines that are
extended in a controlled fashion once the device is landed, thereby
limiting potential damage from abrupt release, and polyurethane
foam surrounding which forms a foamed leading edge when the
constrained device is unsheathed.
The WAVECREST I trial (multicenter, prospective, nonrandomized
registry) recruited 73 patients from Europe, Australia,
and New Zealand, with mean CHADS2 score of 2.5, prior cerebral
embolism in 34%, and a warfarin contraindication in 49%.101 After
TEE-guided deployment, dual antiplatelet therapy was administered
for 90 days and then aspirin continued long-term. Successful
deployment with acute closure was seen in 68/73 (93%), with ≤3mm
peri-device flow at 6 weeks in 65/68 (96%). Acute tamponade
occurred in 2/73 (3%) and there was no procedural stroke, device
embolization or device-related thrombosis.
The device has received CE-Mark in 2013. The pivotal US
WaveCrest II trial is anticipated in 2014.98
With the evidence supporting LAA occlusion for stroke prophylaxis in AF, the increasingly diverse technologies becoming
available for LAA ligation and exclusion, and the parallel development
of medical technologies such as novel anticoagulant agents, there
is an armamentarium of therapeutic options. Coupled with this,
several questions have arisen and remain unanswered, including the
role of LAA ligation when used in conjunction with or in place of
novel anticoagulant agents, if post procedural antiplatelet agents or
anticoagulants are required, the mechanism of recurrent thrombosis
and late appearing leaks, the risk attributable to residual or recurrent
leaks, structural remnants such as beaks, pits, and side lobes, and what
the differences are with ligation versus exclusion.
For now, the current range of products allows for individually tailored
therapy. For example, patients with absolute contraindications for
any anticoagulation, even temporary, an epicardial technique which
does not require adjunctive endocardial access may be better suited,
whilst others with pericardial adhesions would be best served with
an endocardial approach, and some patients may require combined
approaches including with direct surgical visualization.4 There is
evidence that ligation, by silencing the electrical activity of the
appendage,102-104 may provide additional antiarrhythmic benefit in
atrial fibrillation. With increased understanding of the interactions
between device design, appendage anatomy, clinical risk of thrombosis,
and medium to long term success of occlusion, we may recognize how
specific strengths can be harnessed and geared towards the patient at
hand. Ultimately, through better understanding these determinants,
improved device design and deployment technique, and controlled
clinical comparisons of strategies, an ideal closure approach may be
realized.