`
`
`
`
`
`
`
`
`
`
`Publishers, Inc.
`
`
`
`
`
`
`
`
`Jafar Vossoughi
`Nicholas Kipshidze
`John W. Karanian (Eds)
`
`a},ng Medical
`2e;
`and
`Engineering
`
`Edwards Lifesciences Corporation, et al. Exhibit 1115, Page 1 of 24
`
`

`

`
`
`© Medical and Engineering Publishers Inc 2000
`
`First published in the United States of America by
`Medical and Engineering Publishers, Inc
`PO Box 11834
`Washington, DC 20008
`USA
`(http://www.erols.com/medengrgpubinc/)
`
`All rights reserved. No part of this book may be reproduced,
`stored in any retrieval
`system without
`the prior written
`pennission of
`the publisher.
`This includes copying or
`production
`by
`any means, mechanical,
`photographic,
`electronic, or any other means, as well as translation into other
`languages.
`
`l
`
`Library of Congress Cataloging—in-Publication Data
`
`
`
`
`
`I
`
`
`
`
`
`
`Stent Graft Update
`J'
`Edited by Jafar Vossoughi, Nicholas Kipshidze, John W
`F‘
`Karanian. First edition
`
`'\
`Includes bibliographical references and index
`2000
`
`
`LC #2 00—107409
`
`ISBN #: 1-930636-00-8
`
`
`
`
`Copyright 2000
`
`Medical and Engineering Publishers, Inc
`
`
`
`Printed in the United States of America.
`
`Printed on acid-free paper.
`
`
`
`Edwards Lifesciences Corporation, et al. Exhibit 1115, Page 2 of 24
`
`

`

`
`
`The relatively high incidence of restenosis following arterial balloon
`angioplasty and the less frequent, but clinically important complication of
`abrupt vessel closure have directly lead to the development of intraarterial
`stents. A stent is a prosthetic intraluminal scafiolding device designed to
`maintain lumen patency in the setting of an underlying intrinsic or
`extrinsic disease state. Currently, stent placement is being increasingly
`utilized for the treatment of atherosclerotic coronary and peripheral
`vascular disease.
`In the heart, emergency coronary artery stent placement
`is used as a bailout procedure for abrupt or threatened artery closure due
`to arterial dissection following percutaneous
`transluminal coronary
`angioplasty (PTCA) and has reduced the need for urgent coronary bypass
`surgery [27, 33, 58].
`The use of stents as primary therapy for
`atherosclerosis has gained wide acceptance, with reports of reduced
`restenosis rates in selected coronary lesions compared with PTCA [20, 21,
`24, 62].
`
`Keywords: Stent, stent grafl, artery, atherosclerosis, pathology
`
`Renu Virmani, MD (corresponding author), Department of Cardio-
`vascular Pathology, Armed Forces Institute of Pathology, Washington, DC
`20306-6000
`
`The opinions and assertions contained herein are the private views of the
`authors and are not to be construed as ojjicial or reflecting the views of the
`Department of the Army or the Department ofDefense.
`Stent Graft Update, Edited by J Vossoughi, N Kipshidze, JW Karanian
`@2000 Medical and Engineering Publishers, Inc
`(http://www.erols.com/m‘edengrgpubincl)
`
`141
`
`Chapter 10
`
`A PATHOLOGIST‘S VIEW OF STENTING:
`
`BARE STENTS AND STENT GRAFTS
`
`Andrew Farb, MD, Frank D Kolodgie, PhD, Renu Virmani, MD
`
`Edwards Lifesciences Corporation, et al. Exhibit 1115, Page 3 of 24
`
`

`

`been introduced for revascularization of atherosclerotic lesions as well as
`treatment for aneurysms of the aorta and its major branches, traumatic
`aneurysms, and arteriovenous fistulas [29]. With the presence of a graft
`in addition to a stent, it is hoped that there will be a reduction in restenosis
`that occurs secondary to neointimal growth from the surrounding vessels
`[17].
`In the treatment of peripheral aneurySms, the purpose of the stent
`graft
`is to isolate the aneurysmal segment from blood flow thereby
`reducing
`the
`risk of
`aneurysmal
`rupture
`and
`embolization of
`atherothrombotic material from the aneurysm sac [17].
`
`While stents reduce restenosis rates in carefiilly selected lesions [24,
`63], in-stent restenosis remains a recognized clinical problem [31] and can
`be expected to increase in incidence as stenting becomes more frequent
`and is utilized in less ideal lesions [55]. Despite the tremendous expansion
`of the use of stents, there has been little published data on the pathology of
`stents deployed in human atherosclerotic arteries [2, 32, 67]. Most
`experimental studies of stenting in large animals employ normal, non-
`atherosclerotic vessels.
`In these studies, thrombus formation and acute
`inflammation have been identified early after stenting; smooth muscle cells
`and scattered chronic inflammatory cells predominate in the intima 2—4
`weeks
`post-deployment
`[7].
`When
`stents
`have
`been
`used in
`atherosclerotic models (mostly rabbit aorta and iliac arteries), early
`thrombus formation followed by smooth muscle cell proliferation has been
`observed [45, 50, 65]. Studies in normal arteries in experimental models
`suggest a relationship between the severity of arterial
`injury and
`subsequent neointima formation [51, 52, 59].
`
`In this chapter, a general overview of stents and stent grafts will be
`presented.
`The pathology of stenting with and without grafts
`in
`experimental animals and in human atherosclerotic coronary arteries will
`be described to highlight stent-vessel wall interaction.
`
`STENTS
`
`The development of stents was pioneered in the experimental work of
`Dotter (stainless steel and nitinol wire coil) [19], Cragg (nitinol coil stent)
`[14], Mass (self-expanding stainless steel coil stents) [36], Gianturco
`(spring-loaded zigzag stainless steel and interdigitating coil stents) [53,
`74], and Palmaz (stainless steel slotted tube stent)
`[43] which was
`followed by initial stent implants in humans by Sigwart (self-expanding
`mesh stent) [64].
`
`''IIJr
`
`.‘:r_
`
`...H.'
`
`Stainless steel has excellent tensile strength and is
`steel alloys [66].
`resistant to corrosion, but
`is relatively thrombogenic compared to the
`other metals. Titanium has high strength per unit density allowing for
`lighter devices, is corrosion-resistant, and has excellent biocompatibility.
`The shape memory property of nitinol allows for devices to be highly
`compressed and resume original shape upon wamling. Nitinol exerts a
`strong radial expanding force and good biocompatibility. Gold coating of
`stainless steel adds visibility via enhanced radio-opacity and inhibits
`corrosion and infection. The basic designs of stents currently in use are
`the coil, slotted tube, mesh, and corrugated ring.
`
`BIOLOGIC RESPONSES TO STENTING
`
`Stent placement elicits the following sequence of local vascular events:
`thrombus
`formation and inflammation early afier deployment with
`subsequent neointimal growth. These events are affected by changes in
`local shear stress, blood flow velocity, stent design, vessel injury, and the
`presence or absence of underlying pathology in the artery in which the
`stent is deployed. The biologic responses to stent grafis can be expected
`to be similar to those found after placement of bare stents.
`
`Thrombosis
`
`Damage to the vessel wall endothelial lining, either by a bare stent or stent
`graft, induces platelet activation and aggregation accompanied by fibrin
`deposition resulting in thrombus formation. The stent wires themselves
`produce turbulence proximal and distal to the strut resulting in focal low
`Shear stress areas which can precipitate increased thrombus formation.
`Thrombus deposition on stent struts is a uniform finding following device
`deployment, but thrombus size may be affected by stent design [51].
`Tubular slotted stents implanted in normal canine coronary arteries show
`thrombus composed of platelets, fibrin and trapped erythrocytes with an
`immature endothelial
`layer at
`1 week [56].
`In the porcine restenosis
`model, thrombus adjacent to mesh stent struts is composed of platelets,
`fibrin, acute inflammatory cells, and trapped erythrocytes [7].
`In human
`atherosclerotic arteries, stent-associated thrombus is common, particularly
`early after stenting, consisting of platelets, fibrin, and inflammatory cells
`(Figure l) and is seen with all stent designs evaluated (Palmaz-Schatz,
`Gianturco-Roubin, Gianturco—Roubin II, Multi—Link, and Wiktor) [22].
`
`Inflammation
`Stent placement invariably involves injury to the underlying artery by the
`stent itself, that is embedded in the arterial wall, and by adjunctive PTCA,
`
`
`
`Edwards Lifesciences Corporation, et al. Exhibit 1115, Page 4 of 24
`
`

`

`Plaque compression by stent struts is uniformly seen in experimental
`studies and in human arteries [2, 22, 45].
`In animal models, acute
`inflammatory cells
`are observed at 24 hours post-stenting with
`macrophage infiltration by 7 days [7].
`In human atherosclerotic coronary
`arteries, neutrophils accumulate in the first week afier deployment, begin
`to decline in number by 1 month, and then are no longer seen beyond 30
`days [22, 67]. The extent of inflammation increases with greater arterial
`injury and with the penetration of stent struts into the necrotic core of
`atherosclerotic plaques (Figure 2). Chronic inflammation (lymphocytes
`and macrophages) around stent struts is also commonly seen at all time
`points post-stenting and probably reflects the chronic inflammatory
`response already commonly seen in atherosclerotic lesions that is fiirther
`augmented by the presence of the stent acting as a foreign body [22, 30].
`
`
`
`FIGURE 1
`
`A fibrin and platelet-rich thrombus (t, outlined by arrowheads)
`containing inflammatory cells is present adjacent to an AVE stent
`strut (arrow) 8 hours post-deployment.
`A ruptured fibrous cap
`overlying the plaque (p) necrotic core is present below the strut.
`(Hematoxylin-eosin stain)
`
`Cellular Proliferation
`showed
`al
`In the porcine coronary artery restenosis model, Carter et
`maximal cellular proliferation 7 days after stent placement (18.6 :1: 3.5%
`
`
`
`Arterial inflammation in coronary arteries with stents placed 53 days
`antemortem. Movat pentachrome stains are presented in the left
`panels with higher power hematoxylin-eosin from same section shown
`in the right panels.
`In A, few inflammatory cells are present adjacent
`to Palmaz-Schatz strut (*) in contact with fibrous plaque (p).
`In B,
`increased numbers of inflammatory cells are associated with a Palmaz
`Schatz strut (*) that penetrates into a necrotic core (c).
`In C, a
`Palmaz Schatz strut (*) is in contact with damaged media (m) with
`dissection (d) and numerous associated inflammatory cells.
`(A and B
`scale bars 0.10 mm; C scale bar 0.14 mm). Reproduced with
`permission from reference 22.
`
`
`
`Edwards Lifesciences Corporation, et al. Exhibit 1115, Page 5 of 24
`
`

`

`In human stented arteries, cellular proliferation has been evaluated in
`atherectomy specimens from patients with in-stent restenosis.
`In a study
`of 10 peripheral arterial lesions (6 femoral arteries, 3 iliac arteries, and 1
`subclavian artery) with stents implanted 4 to 25 months, lesions from the
`stented
`segments were
`hypercellular with
`actin
`positive
`cells
`demonstrating high proliferative activity (24.6 i 2.3% PCNA-positive).
`In contrast, another study of restenotic tissue obtained by coronary
`atherectomy from stents in place 25-23 months showed a relatively
`hypocellular extracellular matrix; cellular proliferation (determined by Ki—
`67 staining) was rare [1 U]. The reasons for the conflicting results among
`these atherectomy studies are uncertain and are probably due to variability
`and small size of the samples obtained by percutaneous atherectomy.
`
`Stent En dothelialization
`
`experimental animals (rabbit, dog, pig), complete
`In a variety of
`endothelialization of the stent surface may be seen as early as 7 days to
`approximately 4 weeks [28, 45, 56, 69]. Data from humans regarding
`stent endothelialization is limited. Anderson et al
`reported coronary stent
`endothelialization in a single stent 21 days post-implant
`[2]. Van
`Beusekom et al demonstrated complete endothelial stent coverage by 3
`months in saphenous vein bypass grafis [67]. Repair on the luminal
`surface via reendothelialization may be a critical process in vascular
`healing via prevention or limitation of luminal
`thrombosis.
`Further,
`inhibition of continued neointimal growth has been felt to be a result of a
`restored endothelial surface. However, recent data demonstrate that the
`presence of an endothelial surface in and of itself is insufficient to inhibit
`the neointima; normal endothelial
`function and/or non-endothelium—
`dependent factors may be relatively more important than the presence or
`absence of an endothelial surface in suppressing neointimal expansion
`[68].
`
`Neointimal Growth
`The mature
`neointima
`
`consists of
`
`smooth muscle
`
`cells
`
`in
`
`a
`
`demonstrate
`studies
`Experimental
`proteoglycan/collagen matrix.
`organization of the intimal thrombus by 7-14 days with smooth muscle
`cell migration from the media and adventitia [61].
`Cell proliferation
`within the maturing neointima is accompanied by matrix synthesis and the
`formation of a thick cellular neointima by 8 weeks [7, 56, 65].
`In human
`atherosclerotic arteries, neointimal development is delayed compared with
`experimental animals, but is clearly present in all stents implanted 230
`days (Figure 3) [2, 22, 30, 32]. Proteoglycans are macromolecules that
`
`
`
`FIGURE 3
`
`In—stent neointimal (n) growth 16 weeks after placement of a Palmaz-
`Schatz stent in the right coronary artery (panel A). Underlying
`fibrous plaque (p) is present. A higher power view of the neointima
`(panel B) demonstrates the proteoglycan-rich neointima (n) containing
`numerous smooth muscle cells. Near stent struts (*), an organizing
`thrombus (arrowhead) is seen. Fibrous plaque (p) and necrotic core
`(m) are indicated. (Movat pentachrome stain)
`
`are an important component of the extracellular matrix. They consist of
`specific glycosaminoglycans
`(chondroitin sulfate, dermatan sulfate,
`heparan sulfate, keratin sulfate) linked to a protein core via O-gylcosidic
`
`
`
`Edwards Lifesciences Corporation, et al. Exhibit 1115, Page 6 of 24
`
`

`

`Neointimal cellularity and proteoglycans in chronic coronary stents
`compared with balloon angioplasty (PTCA).
`A1: Low power micrograph of left circumflex coronary artery
`containing a Gianturco-Roubin stent placed 10 months antemortem.
`The neointima is outlined by arrowheads, and the stent strut
`identified (*). The neointima is relatively thicker over the stent strut
`compared with the remainder of the arterial segment. Scale bar 0.20
`mm.
`
`A2: Alpha actin staining of neointima (n) identifying smooth muscle
`cells. Scale bar 0.12 mm.
`
`A3: Strong alcian blue stain of the neointima (n) showing presence of
`proteoglycans, predominately of chondroitin sulfate and hyaluronic
`acid. Scale bar 0.12 mm.
`
`B1: Low power micrograph of left anterior descending coronary
`artery treated with PTCA 13 months antemortem. The neointima is
`outlined by arrowheads, and the residual
`lumen (L)
`is
`indicated.
`Scale bar 0.16 mm.
`
`BZ: Alpha actin staining of neointima (n) identifying smooth muscle
`cells. The cell density is similar to the stented artery (A2). Scale bar
`0.20 mm.
`
`B3: Alcian blue stain of the neointima (n) showing strong staining for
`proteoglycans, similar in intensity to the stented artery (A3). Similar
`to stents, chondroitin sulfate and hyaluronic acid are the major
`constituents of the neointima] proteoglycans after PTCA. Scale bar
`0.08 mm.
`
`(Movat pentachrome: A1 and B1; smooth muscle actin immunostain:
`A2 and 32; alcian blue A3 and B3). Modified with permission from
`reference 22.
`
`In our analysis of stented human coronary arteries, the mean
`process.
`neointimal area and neointima area/stent area in successes were 2.2 :1: 1.1
`
`mm2 and 0.39 i 0.12, respectively, versus 3.9 i 1.9 mm2 and 0 68 i 0.15
`in failures,
`respectively (p<0.006 and p<0.0001).
`There were no
`differences between successes and failures with respect to areas of the
`EEL, IEL, plaque, or stent [22]. There was a significant linear correlation
`(p<0.0001, R2=0.54) between increased neointima] growth and increased
`stent size relative to the proximal reference coronary artery lumen [22].
`Therefore, stent over-sizing relative to the reference lumen appears to be
`an undesirable goal in deployment.
`
`
`
`Recently, we demonstrated that similar strong
`human arteries [10].
`immunohjstochemical staining for hyaluronic acid and chondroitin sulfate
`in the neointima of stented coronary arteries compared with PTCA-treated
`coronarv arteries matched for the duration following stent depIOyment or
`PTCA (Figure 4). Neointimal smooth muscle cell density was also similar
`in matched stented and PTCA-treated coronary arteries [22].
`The
`neointima] area was greater in the stented segments compared with PTCA
`arteries (3.06 d: 1.63 mm2 versus 1.94 i 1.20 mm: respectively, p<0.05).
`However, [EL area was larger (11.61 :1: 2.14 mm') with stent placement
`compared with PTCA (2.83 :i: 2.13 mm]. p=0_000|) So that the neointima
`corrected for artery size [neointimaflEU was similar in stents and PTCA.
`
`Ultimate histologic success is dependent on lumen area and neointima
`growth within the stent [22].
`In contrast to PTCA alone, stenting prevents
`negative remodeling of the artery as a compenent of the restenosrs
`u
`
`
`
`
`
`Edwards Lifesciences Corporation, et al. Exhibit 1115, Page 7 of 24
`
`

`

`ranging from superficial damage to the endothelium and superficial layers
`of the media to full
`thickness medial rupture and adventitial stretch.
`Schwatz et al.
`showed the important positive correlation between the
`severity of arterial
`injury in stented porcine coronary arteries and
`subsequent neointimal growth [59]. Other experimental studies suggest
`important
`relationships
`among inflammation, vascular
`injury.
`and
`neointimal growth.
`ln stented nonatherosclerotic balloon—injured rabbit
`iiiac arteries. peak monoeyte adherence was observed 3 days after stenting
`with maximal proliferation seen at 7 days [52]
`There was a linear
`correlation {R2=0.82-0.92) between monocyte adherence and neointirna at
`14 days [52]. Further, increased vascular injury correlated with increased
`neointimal growth, inflammation and thrombus formation [5 I ].
`
`vessel
`underlying
`the
`of
`characteristics
`the
`effect of
`The
`(morphologically normal media verSus absent media due 10 medial tear]
`adjacent to the stent strut was addressed in a porcine double arterial injury
`model [8].
`In areas where the arterial media was absent [damaged by
`balloon angioplasty 4 weeks prior
`to stent placement), neointimal
`thickness was greater dian at stmt sites that were adjacent to an intact
`internal elastic lamina and media [8]
`in stented atherosclerotic human
`coronary arteries, increased arterial injury was associated with increased
`stent-associated inflammation [22].
`Furthermore,
`in human coronary
`arteries containing stents for >30 days, neointimal thickness at stent strut
`sites was greater when medial damage (medial laceration or rupture) was
`present (0.69 i 0.29 mm) compared with struts in contact with plaque
`(0.33 i 0.26 mm, p<0_0001) or struts in contact with an intact media
`[0.29 a 0.23 mm, p<0.0001)
`[22].
`Taken together.
`the data from
`experimental work
`showing
`correlations
`among
`arterial
`injury,
`inflammation and neointima,
`and observations
`from human stents
`demonstrating increased inflammation and neointimal growth when medial
`damage is present suggest
`that a reduction of arterial
`injury during
`catheter-based interventions with stents may have a beneficial effect on
`late neointimal growth. Novel devices that do not require very high
`balloon inflations to accomplish close apposition of the stent to the arterial
`wall (e , self-expanding bare stents) are currently in clinical trial [25].
`
`STENT GRAFTS
`
`Offering the potential advantages of endovascular stents and bypass
`grafis, and deliverable via a percutaneous approach, stent grafts have
`
`been the standard surgical approach for the management of peripheral
`arterial athero-occlusive and aneurysmal disease [1, 3, 54, 70]. However,
`long-term patency of synthetic grafts <6.0 mm in diameter have been
`complicated by thrombosis and intimal hyperplasia at the anastomotic
`sites [57]. The higher rate of graft thrombosis using synthetic material as
`compared with autologous veins may be secondary to incomplete
`endothelialization of the graft [57] as endothelialization >1.5 cm from the
`anastomosis of a synthetic interpositional grafts is rare. Additionally,
`smooth muscle cell proliferation and migration lead to neointimal
`hyperplasia secondary to medial
`injury to the host vessel during the
`creation of the anastomosis itself [60].
`
`Building on the work on Stents by Dotter, Cragg, Balko, and Paroldi
`pioneered the deVelopment of stent grafts for treatment of vascular disease
`and aneurysms
`[4,
`15, 47].
`Currently,
`intraluminally deployed
`endovascular stent grafts (tubular designs with or without bifurcated
`segments) are becoming an alternative for the management of iliofemoral
`occlusive disease [37-39, 48]. These grafts do not require the creation of
`a sutured anastomosis and thus do not cause anastomosis-associated
`vascular injury [72].
`
`In patients and in experimental preparations, autologous veins or
`synthetic materials
`such
`as
`polyester
`(Dacron)
`or
`expanded
`polytetrafluoroethylene (ePTFE) have been used as graft material for
`covered stents [16, 46, 49]. Because it
`is dilatable, PTFE can be
`fashioned as a small profile stent grafi system, delivered to the arterial
`treatment site via a percutaneous sheath, and balloon-dilated to the
`appropriate size [15]. However, because of a 30% elastic recoil, PTFE
`grafts must be over-dilated [15].
`In contrast, Dacron is inelastic and is
`not dilatable in situ [15]. Therefore, precise sizing of Dacron stent grafts
`before deployment is necessary for optimal clinical outcome.
`
`In “unsupported” stent grafts, the stents are located at the ends of the
`graft material and act as tethering points of the stent graft to the arterial
`wall. “Fully-supported” stent grafts, in which the stent is present for the
`entire length of the graft, provide greater longitudinal support and radial
`strength and are less prone to kinking [15]. The stent itself may be on the
`inside,
`the outside, or completely enveloped within the graft material.
`Stent placement inside of the graft allows for enhanced sealing of the
`interface of the graft material with the surrounding vessels
`[15].
`However, this endoskeleton design provides for an uneven flow surface
`
`
`
`Edwards Lifesciences Corporation, et al. Exhibit 1115, Page 8 of 24
`
`

`

`
`
`
`
`local thrombus deposition. Complete interweaving of the stent within the
`graft (re, graft material on both sides of the stent) may be advantageous
`for long—term outcome as poor patency rates have been associated with
`endoprostheses in which the graft is unsupported andfor loosely attached
`to the stent [18, 73]. Further, completely enveloping the stent within the
`body of the ePTFE graft provides a more uniform luminal surface
`compared with a graft placed outside of the stent.
`
`An advantage of covered stent grafts is prevention of embolization of
`friable atherosclerotic plaque material during device deployment. The
`ability for bare stems to penetrate the plaque necrotic core and potentially
`liberate atheroembolic material (Figure 5) was evident
`in our study of
`human coronary stent implants; the lipid core was focally penetrated by
`stent struts in 26% of arterial sections [22]. Plaque prolapse between
`stent struts, especially extrusion of the thrombogenic lipid core, are
`important events that may be effectively prevented via the placement of a
`stent graft.
`
`Endothelialization
`In interposition grafts. healing and endothelialization proceed from the
`anastomotic ends towards the center of the grafi [l2]. Enclothelialization
`appears to be superior
`in cndoluminally placed PTFE stent grafls
`compared to interpositional grafts [41].
`The importance of graft
`proximity to endothelial cells was shown in a study by Bull et al
`in which
`interpositional ePTFE grafts implanted in canine carotid arteries showed
`endothelial cell migration from the native artery extending only 1.0 cm
`beyond the ends of the device [6].
`In contrast, wrapping the external
`surface of the graft with a vein resulted in complete endothelializatioa of
`the ePTFE graft. Endoluminal deployment of a porous stent grafi places
`the device in close proximity to a source of endothelial cells (vessel wall
`lining and vasa vasorum of the surrounding artery) along the entire length
`of the device and may be associated with exposure of the grafi
`to
`physiologically
`important
`local
`concentrations of growth factors,
`cytokines, or other intercellular messengers, such as nitric oxide, which
`can difiuse through the porous graft [60]. These difiusable factors may
`further augment endothelialization in contrast to interposition grafts in
`which endothelialization proceeds from the ends of the grafi only. The
`optimal pore size for graft endothelialization and healing is 60-90 pm
`[26].
`
`
`
`Edwards Lifesciences Corporation, et al. Exhibit 1115, Page 9 of 24
`
`

`

`A 77 year old woman presented with an acute myocardial infarction
`and was treated with direct balloon angioplasty and stenting of the
`right coronary artery. Refractory myocardial no-reflow developed
`resulting in cardiogenic shock and cardiac arrest.
`In the stented
`portion of the right coronary artery, there was focal penetration by
`multiple stent struts into a large hemorrhagic necrotic core (nc, panel
`A). Several penetrating and non-penetrating stent strutsjare indicated
`(*). At higher power (panel B), a penetrating strut (*) is better
`visualized within the necrotic core (nc) that is rich in cholesterol clefts.
`The end (arrowhead) of the ruptured fibrous cap (fc)
`is present.
`Numerous small intramyocardial coronary arteries were occluded by
`atheroemboli containing cholesterol clefts (cc's, panel C) that were
`responsible for the myocardial no-reflow. A potential advantage of
`stent
`grafts
`over
`bare
`stents
`is
`a
`reduced
`frequency
`of
`atheroembolism.
`(Movat pentachrome: A and B; hematoxylin—eosin:
`C)
`
`Inflammation
`in experimental animals produces acute
`Dacron stent grafi placement
`reaction associated with the graft and
`inflammation and a giant cell
`extending into the adjacent vessel wall
`[34]. Clinically, this inflammatory
`response has been postulated to correspond to the perivascular thickening
`observed by magnetic resonance imaging that lasts 4-6 weeks [35] and a
`"post-implantation" syndrome characterized by fever,
`leukocytosis, and
`elevation in serum C-reactive protein [5, 17].
`In human endovascular
`stent graft explants, a foreign body type reaction was seen when the
`device had an external wrap on the PTFE or if placed within the adventitia
`[38].
`
`Neointimalformation, endothelialization, andfinal healing
`In interposition grafts, neointimal
`thickening occurs at
`the surgical
`anastomosis as a result of smooth muscle cell migration and proliferation
`secondary to medial injury in the host artery [11, 13, 23, 40]. Reduced
`neointimal growth is potential advantage of endoluminally placed grafts;
`Ombrellaro et
`al
`found that
`intimal hyperplasia was greater
`in
`conventionally placed grafts as compared to endovascular grafis [42].
`However, it must be recognized that neointimal growth along the length of
`a intraluminally placed stent graft will be proportional
`to the arterial
`injury associated with deployment. The use of self-expanding stent grafts,
`rather than balloon-expandable stents, may result in less arterial injury
`[9]. Marin described histopathologic findings
`from 7 stent grafis
`
`e'ndothelium was present in the perianastomotic region at 6 weeks and
`extended 1-3 cm from the ends of the graft at 3 months [3 8]. There was
`little neointimal development within the graft
`[33].
`A subsequent
`angrographic study demonStrated less neointimal growth in the PTFE-
`covered'portion of a Palmaz stent grafi versus the uncovered portion of
`the devrce, suggesting that the presence of PTFE may inhibit smooth
`[[13133ch cell migration into and proliferation within the graft neointima
`
`HEMOBAHN STENT GRA FT
`
`We recently completed an evaluation of the long-term patency and healing
`characteristics of the HEMOBM—II‘ITM stent graft in normal canine iliac
`and femoral arteries [71]. This device consists of a nitinol stent lined with
`an ultrathin ePTFE material. The thickness of the ePTFE used in this
`devrce (lOO-pm) is approximately five times thinner than the standard
`commercially available ePTFE used in surgically placed vascular grafis.
`The intemodal distance (between two solid nodes of PTFE) is 30 um, and
`a nonporous FEP/ePTFE laminate (1.0 mm width) bonds the graft to the
`stent along its entire length with 1.0 mm gaps so that half of the graft
`surface area remains porous.
`Further, unlike other endoprostheses in
`which the graft is sutured only to the ends of the stent [18], the ePTFE in
`thrs devrce is completely interweaved with the stent providing longitudinal
`support.
`
`these stent grafts were placed with fluoroseopic and
`In this study,
`.
`rntravascular ultrasound guidance In canine iliofemoral arteries, and stent
`grafts were analyzed at two weeks, one month, three months, six months
`and twelve months post-deployment.
`All stent graft
`implants were
`successfully deployed in the desired locations without foreshortening. All
`devrces were widely patent at all
`time points.
`Two weeks after
`implantation, there was fibrin deposition covering 70-80% of the luminal
`surface and an inflammatory infiltrate near the graft material and the
`surrounding nitinol wires (Figure 6). At one month, there Was increased
`fibrin deposition between the endoprostheses and the native artery the
`inflammatory infiltrate was minimal, and no giant cell
`reaction ywas
`evrdent. Neointimal thickness in the mid-region of the device was less at 1
`month (0.42 :l: 0.02 mm) than at 2 weeks (065 :I: 0.07 mm), most likelv
`secondary to organization of mural
`thrombus.
`Three months after
`deployment, there was nearly complete healing of the space between the
`
`
`
`Edwards Lifesciences Corporation, et al. Exhibit 1115, Page 10 of 24
`
`

`

`
`
`there was focal
`of late neointimal growth, was mild with this device;
`medial compression but no medial
`laceration or rupture.
`The self-
`expanding property of the nitinol stent used,
`in contrast to balloon-
`expanded stent grafis, may be an important feature in limiting damage to
`the underlying artery. Overall, the pattern of pathological responses--
`early mural
`thrombus
`associated with
`inflammation followed by
`neointimal growth--is similar to that observed in bare stenting in human
`atherosclerotic arteries.
`
`FIGURE 6
`
`FIGURE 7
`
`Hemobahn ePTFE-nitinol stent graft placed in canine iliofemoral
`arteries and evaluated 2 weeks post—deployment. The stent wires are
`enveloped with the ePTFE layers. A thin layer of thrombus (t) is seen
`(panel A) on the luminal surface of the stent graft. Between the stent
`graft and the media of the native artery, a thin layer of inflammatory
`cells (ic‘s) is present (panel B). (Hematoxylin-eosin stain)
`
`ePTFE and the media with smooth muscle cells in a proteoglycan matrix.
`At
`three and six months after
`implantation, neointimal giant cells,
`thrombi, and calcification were observed in a few cases. At one year,
`macrophages were rare, and there were no giant cells. Neointimal growth
`(Figure 7) was observed in all grafis with a mean thickness of only
`0.42 i 0.03 mm at 12 months, determined by histomorphometry, resulting
`in only a 6% reduction in luminal area (measured by intravascular
`
`Hemobahn stent graft placed in canine iliofemoral arteries and
`evaluated 6 months post-deployment.
`The neointima (n)
`is fully
`organized consisting mostly of SMCs in a proteoglycan and collagen-
`rich matrix. A thin layer of neointima is also present between the
`outer layer of the graft and the media (m). (Hematoxylin-eosin stain)
`
`Endothelialization was assessed by scanning electron microscopy. At 3
`months, 75% of the surface of the graft was covered by endothelial cells,
`and at 6 and 12 months,
`there was 90-99% coverage (Figure 8).
`Endothelial cells were predominately spindle shaped and oriented in the
`direction of flow. There were no adherent platelets or inflammatory cells
`on the endothelial surface, suggestive of a functional endothelial lining.
`previous studies, approximately two and a half times thiCker ePTFE lined
`
`
`
`Edwards Lifesciences Corporation, et al. Exhibit 1115, Page 11 of 24
`
`

`

`
`
`FIGURE 8
`
`Scanning electron micrographs of an ePTFE-nitinol endoprosthesis
`explanted at 6 months; the device is opened longitudinally exposing
`the luminal surface (panels A-C). The device is 90% endothelialized
`(panel A). The distal end of the stent graft (bottom of panel A) shows
`a
`focal area of non-endothelialization (arrowheads).
`At high
`magnification in the mid-portion of the graft (panel B), there is near
`complete endothelialization.
`At higher magnification (panel C),
`longitudinally oriented endothelial
`cells with tight
`intercellular
`junctions are evident.
`
`
`
`endothelialization only at the end stent grafts