From the Society for Vascular Surgery

FromTan

Thede

In mAuthPresSu

AddRepM(e

Thetom

0741Cophttp

164

Histologic analysis of stent graft oversizing in thethoracic aortaIgor Rafael Sincos, MD, PhD,a,b Erasmo Simão da Silva, MD, PhD,a,b Sergio Quilici Belczak, MD, PhD,a,b

Anna Paula Weinhardt Baptista Sincos, MD,a Maria de Lourdes Higuchi, MD, PhD,a,b

Vitor Gornati, MD,b José Piñatas Otoch, MD, PhD,a,b and Ricardo Aun, MD, PhD,a,b São Paulo, Brazil

Objective: To elucidate the histologic changes after stent graft oversizing in nonatherosclerotic aortas using an experi-mental porcine model. We previously reported that the diameter and angulation of the aorta in this model are similar tothose in young individuals who undergo stent graft repair for blunt aortic injuries. The lack of commercially availablestent grafts specific for repairing blunt aortic injuries, particularly for small and angulated aortas, may be related to thehigh rate of endograft complications in this population.Methods: Twenty-five pigs were randomized into one control group (without stent graft implantation) and four oversizedgroups (A: 10%-19%, B: 20%-29%, C: 30%-39%, and D: >40%). Three circumferential fragments were collected from theaorta for histologic and immunohistochemical studies. Morphometric analyzes were performed using an inflow systemand image analysis software (Quantimet 500; Leica Cambridge Ltd, Cambridge, UK).Results: Collagen expression in the aortic wall was not significantly different among the five groups (P [ .5604). Therewere significantly fewer muscle fibers in the aortic wall in the oversized groups compared with the control group(P [ .000198). The proportion of elastic fibers in the aortic wall was significantly smaller in the oversized groupscompared with the control group (P [ .0000001). Immunohistochemical analysis showed that a-actin expression in theaortic wall was significantly decreased in the oversized groups comparedwith the control group (P[ .002031). There wereno significant differences in either the number of muscle fibers or a-actin expression among the four oversized groups.Conclusions:Histologic and immunohistochemical studies confirmed the structural disarrangement of the aortic wall afterinsertion of an endoprosthesis, including reduced number of muscle and elastic fibers. (J Vasc Surg 2013;58:1644-51.)

Clinical Relevance: Blunt aortic injury is a catastrophic event that most commonly affects the descending thoracic aorta,and it is the second most common cause of death in trauma patients. Thoracic endovascular aortic repair is increasinglybeing used to treat such conditions. The prostheses available at the time we conducted the present study were notspecifically designed to treat thoracic aortic trauma. Furthermore, few stent grafts had been designed for use in normal orhealthy aortic tissue or in young patients with acute angulation of the thoracic aorta. To our knowledge, no large studyhas evaluated the optimal target level of oversizing in thoracic aortas, especially for nonatherosclerotic aortas. We haveconducted an experimental study in which we evaluated the biomechanical changes associated with four levels of stentgraft oversizing on a nonatherosclerotic aortic wall. In the present report, we describe the histologic and immunohis-tochemical results of that study.

Blunt aortic injury is a catastrophic event that mostcommonly affects the descending thoracic aorta, and it isthe second most common cause of death in traumapatients.1,2 Traumatic aortic transection, ruptured de-scending thoracic aneurysm, and acute complicated type B

Hospital Israelita Albert Einstein, Centro de Experimentação ereinamento em Cirurgiaa and the Department of Surgical Techniqued Experimental Surgery,b University of São Paulo Medical School.study was sponsored by the Fundação de Amparo à Pesquisa do EstadoSão Paulo (FAPESP) foundation.emoriam: Luiz Francisco Poli de Figueiredo, MD, PhD.or conflict of interest: none.ented at the 2013 Vascular Annual Meeting of the Society for Vascularrgery, San Francisco, Calif, May 30-June 1, 2013.itional material for this article may be found online at www.jvascsurg.org.rint requests: Igor Rafael Sincos, MD, PhD, Endovascular SP Clinic,ato Grosso Street 306 conj 609, São Paulo/SP, Brazil, 05412-002-mail: igorsincos@usp.br or drigor@endovascularsp.com.br).editors and reviewers of this article have no relevant financial relationshipsdisclose per the JVS policy that requires reviewers to decline review of anyanuscript for which they may have a conflict of interest.-5214/$36.00yright � 2013 by the Society for Vascular Surgery.://dx.doi.org/10.1016/j.jvs.2013.02.005

4

dissection are acute pathologies of the descending aortaassociated with a high mortality rate after traditional openrepair. Thoracic endovascular aortic repair is increasinglybeing used to treat such conditions in the descendingthoracic aorta in both elective and emergent surgery.3,4

However, a consensus on the management of blunt aorticinjuries has not been reached because of the very lowquality of evidence.3,5

Patients experiencing traumatic aortic injury are usuallyyounger and have a smaller aortic diameter (19 mm) thanthose with aneurysms.2,5-8 The lack of commercially avail-able stent grafts specific for repairing blunt aortic injuries,particularly for small and angulated aortas, may be relatedto high rates of endograft complications in this popula-tion.2,5,7 A previous study on aortic trauma, published in2008, concluded that improvements in endovasculardevices for blunt aortic injuries were urgently needed.Although that study reported a significant improvementin the survival of patients, 20% of patients developedcomplications related to the implant stent.9 Some of theendograft complications in these patients include endoleak,stroke, and graft collapse.10-16 The prostheses available at

JOURNAL OF VASCULAR SURGERYVolume 58, Number 6 Sincos et al 1645

the time we conducted the present study were not specifi-cally designed to treat thoracic aortic trauma. Furthermore,few stent grafts had been designed for use in normal orhealthy aortic tissue or in young patients with acute angu-lation of the thoracic aorta. This can lead to poor apposi-tion of the prosthesis in the aortic wall, a situation that isaggravated by excessive oversizing.14 Bandorski et al17 re-ported three cases of aortic collapse after treatment witha stent graft and concluded that smaller diameter of theaorta in young patients and excessive oversizing are predic-tors of this complication.

To our knowledge, however, no large study has evalu-ated the optimal target level of oversizing in thoracicaortas, especially for nonatherosclerotic aortas. An interna-tional consensus suggests that oversizing of >20% shouldnot be used following trauma in young individuals becausethe prosthesis can collapse and obstruct the aorta.18 Wehave conducted an experimental study in which we evalu-ated the biomechanical changes associated with four levelsof stent graft oversizing on a nonatherosclerotic aorticwall.2 In the present report, we describe the histologicand immunohistochemical results of that study.

Fig 1. A, Locations of the circumferential fragments. B, Pairedtissue fragments were used for the biomechanical and histologicstudies.

METHODS

The complete study protocol has been publishedbefore.2 This study was approved by the local institutionalreview board (University of São Paulo Medical School) andfollowed the guidelines of the National Health Council ofBrazil. Twenty-five large white pigs weighing 38-50 kg(mean, 41.2 kg) and aged 5-6 months were used in thestudy. The mean aortic diameter close to the left subclavianartery was 20.6 mm (range, 18-24 mm). The pigs wererandomly divided into five groups with five animals pergroup. One group was set as the control group (n ¼ 5),in which no endoprosthesis was implanted, to assess histo-pathologic and morphologic changes of the aortic wallassociated with normal growth. The other four groupsunderwent stent graft implantation (Valiant; MedtronicVascular, Santa Rosa, Calif) with device oversizing of10%-19% (group A), 20%-29% (group B), 30%-39%(group C), or >40% (group D).

At 30 days after stent graft implantation, the animalsunderwent angiography, the aortas were resected forbiomechanical and histological testing, and the animalswere euthanized. The animals were maintained for 30days because this was sufficient time for incorporation ofthe stent graft but was short enough to avoid excessiveanimal weight gain and marked increases in the diameterof the thoracic aorta.19 The surgical protocol, includingimplantation of the stent graft, and the biomechanicaltest were described in our previous report.2 For thepurpose of this study, we collected three circumferentialfragments of the aorta: proximal (X, juxta-subclavian),middle (Y, at 5 cm from the subclavian), and distal (Z, at10 cm from the left subclavian artery). All three fragmentswere in contact with the prosthesis and were paired withthe fragments used in the biomechanical test (Fig 1).

All specimens collected for histology were fixed in 10%buffered formalin. The specimens were then processed,embedded in paraffin, and cut to 5 mm thick sections.The sections were stained with hematoxylin and eosin,Van Gieson, Picrosirius, and trichrome Masson to identifyelastic and muscle fibers, collagen, inflammatory infiltrate,and neovascularization. Stained sections were analyzed bylight microscopy with polarized light. Morphometricanalyzes were performed using an inflow system and imageanalysis software (Quantimet 500; Leica Cambridge Ltd,Cambridge, UK) coupled to an optical microscope witha magnification of �20 and a video camera (3CCD Donpi-sha; Sony, Tokyo, Japan). The mean thickness and totalareas of the intima and the media were determined aftermanually delineating the layers in sections stained withVan Gieson, which stains elastic fibers and facilitates delin-eation of the aortic layers. The abundance of muscle fibers,collagen, and fibrosis was determined by automated imageanalysis of color in sections stained with Masson and Picro-sirius, which are used to diagnose fibrosis or neointimaformation.

The image analysis system was periodically calib-rated using a stage micrometer and the computer was

Fig 2. Cross-section of the aorta (Masson stained). The photo-micrograph shows indentations in the wall of the aorta withneointima and thrombi caused by compression of the metalsupports of the stent. A, View of the entire cross-section(magnification, �1.2). Black arrow-I, intima; red arrow-A,adventitia; M, medial layer; T, mural thrombi. B, Higher magni-fication view (�4). Blue arrow, site of compression; black arrow,gap between the metal struts of the stent.

JOURNAL OF VASCULAR SURGERY1646 Sincos et al December 2013

programmed for a magnification of �20 at the start of eachwork session. For each section, the most representativemicroscopic field was identified and the following measure-ments were performed: (1) digitization of the imageobtained from the selected microscopic field; (2) measure-ment of the areas representing collagen, muscle fiber, orelastic fiber in the digitized image; (3) measurement ofthe surface area of the entire aortic field; and (4) calculationof the relative abundance of each fiber in the aortic field(ie, fiber surface area/total surface area).

To determine the relative abundances of collagen,muscle fibers, and elastic fibers, the percentages of eachfield of view stained blue (collagen), red (muscle fibers),or black (elastic fibers) were automatically calculated usingQuantimet 500 software (Leica Cambridge Ltd). Thehistologic features were separately evaluated in the threelayers of the aortic wall: the inner third (close to theintima), middle third, and the outer third (near the adven-titia). The values for each layer were summed andcompared between each group and with the control group.We also compared these three layers within each group toidentify the layer of the aortic wall that showed the greateststructural change.

In all of the regions covered by the endoprosthesis,compression of the metal supports of the stents leavesindentations in the aorta wall. Thrombi develop in theseindentations, resulting in a sinusoidal pattern on theluminal surface. The abundance of collagen, elastic fibers,and muscle fibers was measured in the areas compressedby the stent and in the uncompressed areas. For thepurpose of this study, we first focused on uncompressedregions to minimize possible bias because the histologicfeatures are more difficult to evaluate in the compressedregions (Fig 2). Additionally, the structures are more accu-rately identified both visually and using Quantimet 500software (Leica Cambridge Ltd) in uncompressed regions.Finally, we compared the histologic characteristics of thecompressed and uncompressed regions to determinewhether the pressure exerted by the stent can modify thehistologic features of the aortic wall because the pressureexerted by the stent on the aortic wall is an importantfactor that may influence integration of the stent into theaortic wall.

Immunohistochemistry. We performed immunohis-tochemistry of a-smooth muscle actin (a-actin) to iden-tify muscle cells and myofibroblasts within the bloodvessels and endothelium, as described by Reis et al20

and Koike et al.21 Tissue sections were incubated over-night with mouse anti-human a-actin antibody (1:500dilution; Sigma Chemical, St. Louis, Mo) at 4�C. Thesections were then incubated with streptavidin-biotinperoxidase (Dako LSABþ kit; Dako, Carpinteria, Calif)and diaminobenzidine tetrahydrochloride for colordevelopment. Finally, the sections were weakly counter-stained with hematoxylin. Vascular smooth muscle cellswith intense staining (brown cytoplasm) were defined aspositive for a-actin.

Terminal deoxynucleotidyl transferase dUTP nickend labeling staining. Cellular apoptosis was detected insitu using the terminal deoxynucleotidyl transferasedUTP nick end labeling (TUNEL) method, as previouslydescribed.22,23 Briefly, 5 mm thick tissue sections wereembedded in paraffin and fixed in formalin. The sectionswere then dewaxed in xylene, rehydrated, and pretreatedwith proteinase K (co-Gib, Gaithensburg, Md) at a dilutionof 20 mg/mL in 10 mM Tris/HCl (pH 7.5) for15 minutes at a temperature of 32�C-35�C. The sectionswere washed in 0.1 M phosphate-buffered saline, pH 7.3,and incubated in humid atmosphere at 37�C for 60minutes in 50 mL of fluorescein for the coupling reaction(Boehringer, Mannheim, Germany). Sections of the aortawith high numbers of apoptotic interstitial cells wereidentified as a positive control. The number of apoptotic(ie, TUNEL-stained) cells in each field was counted at �20and scored as follows: 0 ¼ no apoptotic cells; þ ¼ 1-4apoptotic cells; þþ ¼ 5-10 apoptotic cells; and þþþ $10apoptotic cells.

Statistical analysis. Quantitative data are expressed asthe arithmetic mean and standard deviation, with 95%confidence intervals for the mean. For graphic representa-tion, we plotted the arithmetic mean and 95% confidence

Table II. Relative collagen content in the aortic wall (thepercentage of field of view stained blue in uncompressedareas)‒Masson stained

Group Collagen, mean SD

A 18.15 13.18B 15.14 8.73C 12.69 7.94D 13.89 9.23Control 13.84 6.20

SD, Standard deviation.Values are presented as the mean and SD for the percentage of collagenexpression.

Table I. Comparison of preoperative and postoperativediameters of the aorta

Diameters, mmPreoperative

Average 6 SDPostoperativeAverage 6 SD P value

Pre-subclavian 21.7 6 1.84 22.17 6 1.85 .01366a

Juxta-subclavian 20.6 6 1.46 20.71 6 1.51 .43224Post 5 cm 18.04 6 1.4 17.17 6 1.28 .00224a

Post 10 cm 17.57 6 1.23 16.71 6 1.16 .00156a

SD, Standard deviation.Values are presented as the mean and SD. P values were determined usingStudent t-test.

JOURNAL OF VASCULAR SURGERYVolume 58, Number 6 Sincos et al 1647

interval. Other qualitative data, including proportion ofTUNEL-positive cells, are shown as percentages.

For continuous variables, the variations in mean valueswere compared among the five groups using analysis ofvariance, followed by the post hoc Newman-Keuls test toidentify differences between pairs of groups. Differenceswere confirmed using the nonparametric Kruskal-Wallistest. In all analyses, differences were considered statisticallysignificant at P < .05. Statistical analyzes were performedusing Statistica for Windows software (v. 8.0; StatSoftInc, Tulsa, Okla).

RESULTS

All animals survived the surgical procedures. Therewere no vascular access complications, but one woundneeded to be re-sutured because of an infection that wastreated with antibiotics. This animal recovered and waswell until sacrifice. Mean weight gain was 1.2 kg (range,41.3-42.4 kg). Animal weight and the preoperative andpostoperative diameters of the aorta just after the subcla-vian artery were not significantly different among the fivegroups. These diameters were measured by angiography.However, the diameter of the aorta just before the subcla-vian artery had increased at 30 days after implantation(Table I), probably reflecting normal growth (P ¼.013662). By contrast, the diameter of the aorta measuredat 5 and 10 cm after the subclavian artery had decreased at30 days and was probably related to neointima formation.Postoperative angiography revealed no leak or distal migra-tion of the endoprosthesis.

Collagen. The amount of collagen in the aortic walltended to be higher in groups A and B compared withthe control group (Table II), but no significant differ-ences were found between the groups by analysis of vari-ance (P ¼ .5604).

Muscle fibers. The abundance of muscle fibers in theaortic wall gradually and significantly decreased comparedwith that in the control group (P ¼ .000198; Table III).Although the amount of muscle fiber tended to decreaselinearly with increasing oversizing, it was not significantlydifferent between groups A, B, C, and D (Fig 3). Repre-sentative tissue sections stained for collagen and musclefiber are shown in Fig 4.

Our secondary evaluation of the three thirds of theaortic wall revealed a greater decrease in smooth musclefibers in the inner third of the aortic wall in the tissueregions that were compressed by the stent of the tissuethan in the uncompressed regions (P ¼ .006145). Therewere no significant differences in the other middle or outerthirds.

Elastic fibers. Photomicrographs of histologic sectionsshowing the abundance of elastic fiber in the control groupand group D are shown in Fig 5, online only. Theproportion of elastic fiber in the aortic wall was significantlylower in the four oversized groups compared with thecontrol group (P ¼ .0000001; Table IV and Fig 6, onlineonly). However, there were no significant differencesamong groups A, B, C, and D.

The abundance of elastic fibers in the aortic wall wassignificantly lower in the inner third of the aortic wall inthe tissue regions that were compressed by the stent thanin the uncompressed regions of the tissue (P ¼ .025190).Therefore, the distribution of elastic fibers was similar tothat of muscle fibers.

In our analyses of histologic sections of the aortas fromgroup D (oversizing >40%), we found an image suggestiveof the presence of an aneurysm in the aortic wall. Histo-logic examination showed evidence of disruption of thefibers within the aortic wall, with a marked decrease inmuscle fiber content, and the formation of a muralthrombus (Fig 7, online only).

a-Actin. Immunohistochemical staining of a-actin wasperformed to identify smooth muscle and myofibroblasts inthe aortic wall. Representative a-actin-stained tissuesections are shown in Fig 8, online only, for aortas fromthe control group and group D. The expression of a-actinin the aortic wall was significantly weaker in all four over-sizing groups compared with the control group (Table V).Similar to the evaluation of Masson-stained muscle fibers,we found a tendency for a-actin expression to decreaselinearly with increasing oversizing, but there were nosignificant differences in a-actin expression among the fouroversizing groups.

TUNEL. TUNEL staining, in which apoptotic cellsare stained with green fluorescence, was performed todetect in situ cell death. As shown in Table VI, themajority of cells counted in each section were notapoptotic. Only tissue sections from group B contained

Table III. Relative muscle fiber content in the aorticwall (the percentage of field of view stained red inuncompressed areas)‒Masson stained

Group Muscle fibers, mean SD

A 31.74 17.20a

B 29.41 13.57a

C 27.27 8.66a

D 23.25 7.95b

Control 46.18 16.80

SD, Standard deviation.Values are presented as the mean and SD for the percentage of muscle fiberexpression.aP < .01.bP < .001 vs control. MEAN +

C.I.95%CT A* B* C* D**GROUP

15

20

25

30

35

40

45

50

55

60

MU

SCLE

FIB

ER (

% )

Fig 3. Comparison of relative muscle fiber content in the aorticwall among the five groups (the percentage of field of view stainedred in uncompressed areas)‒Masson stained. Values are presentedas the mean 6 95% confidence interval. *P < .01 and **P < .001vs control.

JOURNAL OF VASCULAR SURGERY1648 Sincos et al December 2013

>10 cells apoptotic cells per field (Fig 9, online only).The control group had the least apoptotic cells.

DISCUSSION

No previous study has examined the effects of oversiz-ing of a thoracic endoprosthesis on structural changes ofthe aortic wall. In our study, we used an experimentalmodel that has a thoracic aorta diameter similar to thatof young adults (19 mm), the most frequent victims ofmotor vehicle accidents, and who consequently havea higher risk of aortic injury.2,3 We tried to reproduce, ina controlled manner, the implantation of a stent graft inthe thoracic aorta with four different levels of oversizing.We then evaluated the histologic and biomechanicalchanges of the aorta.

Our previously reported biomechanical tests showedthat the aortic wall undergoes marked structural changesfollowing stent implantation. Regardless of the level ofoversizing, the aortic wall suffers a considerable loss of elas-ticity, which was biomechanically characterized by strainand maximum deformation. Meanwhile, the maximumstrength, the maximum tension, and the maximum stresssupported by the fragments of the aortic wall suffereda linear and progressive loss with increasing oversizing.The biomechanical tests were performed in three differentregions of the thoracic aorta: close to the left subclavianartery, and 5 and 10 cm after the subclavian artery. Theaortic fragments obtained for the biomechanical testswere divided and paired for histologic study, as illustratedin Fig 1.

The histopathologic results presented here are similarto the previously reported biomechanical results, showingthat greater oversizing of the stent graft may lead to greaterinjury to the aortic wall. The strength of the aortic wall isdetermined by the relative quantity and organization ofcollagen and elastic fibers, which are distributed in a helicalmanner around the circumference of the aorta. The rupturestrength of the wall shows similar behaviors in both thetransverse and longitudinal arrangements of the aortabecause of the interweaved collagen and elastic fibers,although the behavior changes at very high pressures.24

Collagen deposition in the middle layer is a processinvolved in the repair of the aortic wall following injury.25

Although we found no significant quantitative differencesin the number of collagen fibers between each group,our subjective evaluation suggested that collagen deposi-tion was greater in the regions of the wall exhibiting thegreatest damage. These changes in collagen depositionwere accompanied by the loss of elastic fibers and disorga-nization of muscle fibers. Increasing the amount ofcollagen but decreasing elastin and muscular fibers changesthe structural morphology of the aortic wall and results ina structurally weak tissue. We observed clear qualitative andorganizational change in the structure of the aortic wallwith increasing oversizing.

Acute inflammatory cells in the myocardium werereported to be activated and to proliferate in response tobradykinin stimulation. These cells synthesize growthfactors, such as transforming growth factor-b and basicfibroblast growth factor, which promote the phenotypicdifferentiation of cardiac fibroblasts into cells expressinga-actin. The resulting myofibroblasts were thought to bethe main cardiac cell type responsible for collagen synthesisduring wound healing.21

In vascular vessels, activated myofibroblasts andsmooth muscle cells both express several markers includinga-actin, SM22, caldesmon, vinculin, and h-calponin.26

Myofibroblasts can potentially differentiate from circu-lating bone marrow-derived precursors and from the trans-differentiation of SMCs resident in the tunica media orendothelial cells. The expression of a-actin renders fibro-blasts as highly contractile and is a characteristic of myofi-broblast differentiation.26

Myofibroblast differentiation is linked to restenosisafter percutaneous transluminal angioplasty. Restenosis isan excessive wound healing reaction of the vessel subjectedto revascularization and leads to a new narrowing of the

Fig 4. Masson-stained tissue sections of an aorta in the control group vs group A. Blue, collagen; red, muscle fiber. A,Cross-section of the aortic wall in control group (magnification, �4). B, Higher magnification view of the middle thirdof the aortic wall in control group (magnification, �20). C, Cross-section of the aortic wall showed increased collagenand decreased muscle fiber in the middle layer in group A (magnification, �4). D, Higher magnification view of theinner third of the aortic wall in group A (magnification, �8.6).

JOURNAL OF VASCULAR SURGERYVolume 58, Number 6 Sincos et al 1649

vascular lumen. Factors that contribute to vascular resteno-sis include constrictive remodeling, neointima hyperplasia,and neoadventitia formation.21,26

To our knowledge, no prior studies were designed toanalyze the expression of a-actin after stent graft implanta-tion. In our study, a-actin immunohistochemistry was per-formed to identify muscle cells and myofibroblasts withinthe blood vessels and endothelium. The abundance ofmuscle fibers and a-actin expression tended to decreasewith increasing oversizing, which may be related to thereduced strength of the aortic wall observed in biomechan-ical tests. Although muscle cells are not responsible for sup-porting the aortic wall, they are involved in repairing thedamaged wall. The histologic and immunohistochemicalchanges observed here may be related to a deficiency inthe repair mechanism of the aortic wall after stent graftimplantation, which seems to persist for 30 days.

Histologic evaluation of the amount of elastic fiber inthe aortic wall was consistent with the results of the biome-chanical test. We found a considerable loss of elastic fiber inthe aortic wall after stent graft implantation that was inde-pendent of the level of oversizing and was coupled witha loss of biomechanical elasticity, as characterized by strainand maximum deformation. Subjective assessment revealedthat the elastic and collagen fibers had lost their wavy shapein the stent graft groups as compared with the controlgroup. Furthermore, the identification of an aneurysm ofthe aortic wall in the group with greatest oversizing(ie, group D, >40%) is a strong indication that the quanti-tative decrease in elastic fiber content and abnormal

interactions between collagen and elastic fiber change theresistance of the aortic wall to biomechanical stress.

Interestingly, the TUNEL analysis showed no signifi-cant difference in the number of apoptotic cells betweeneach group, although there were fewer apoptotic cells inthe control group than in the other groups. However, itmust be acknowledged that this technique representsapoptosis occurring throughout the duration of the study,and we cannot verify precisely when cellular injury (ie,apoptosis) occurred.

Scheumann et al27 reported medial thinning andluminal medial necrosis in areas covered by a stent graftat 5 days after thoracic endografting in six pigs. Theirresults suggest an early phase of vulnerability to rupturebefore scarring and adaptive changes can strengthen theresidual aorta. By comparison, we found no evidence ofmedial necrosis or inflammation at 30 days after thoracicendografting.

The hypothetical mechanism by which an endopros-thesis causes injury to the aortic wall may involve transmu-ral ischemia caused by increased pressure and obstructionof blood flow. Similar to humans, the aorta of pigs dependson the infusion of nutrients from the vasa vasorum and viaendoluminal routes.28 Implantation of a graft in thethoracic aorta interrupts intraluminal perfusion pressureand occludes collateral circulation from the aortic wall(ie, vasa vasorum). Additionally, the radial pressure exertedby the metal framework can induce focal wall ischemia.Our histologic analyses revealed that the inner third ofthe aortic wall that experienced the greatest compression

Table IV. Relative elastic fiber content in the aortic wall(the percentage of field of view stained black inuncompressed areas)‒Van Gieson-stained

Group Elastic fiber, mean SD

A 23.61 10.15a

B 25.46 8.5a

C 25.87 12.07a

D 22.33 6.48a

Control 55.06 10.82

SD, Standard deviation.Values are presented as the mean and SD for the percentage of elastic fibercontent.aP < .001 vs control.

Table V. Relative -actin expression in the aortic wall (thepercentage of field of view stained brown inuncompressed areas)‒a-actin-stained.

Group a-Actin, mean SD

A 18.62 9.29a

B 19.29 7.16a

C 16.40 8.23a

D 17.26 5.86a

Control 29.70 2.76

SD, Standard deviation.Values are presented as the mean and SD for the percentage of a-actinexpression.aP < .05 vs control.

Table VI. Comparison of the proportion of apoptoticcells in the aortic wall among the five groups

Group 0 þ þþ þþþ

A 86,67% 13,33% 0% 0%B 66,67% 13,33% 6,67% 13,33%C 80% 13,33% 6,67% 0%D 80% 20% 0% 0%Control 92,31% 7,69% 0% 0%

0 ¼ no apoptotic cells; þ ¼ 1-4 apoptotic cells; þþ ¼ 5-10 apoptoticcells; þþþ $10 apoptotic cells.

JOURNAL OF VASCULAR SURGERY1650 Sincos et al December 2013

from the stent had significantly reduced muscle fiber andelastic fiber compared with corresponding tissue that wasnot compressed by the stent. However, in the middleand outer thirds, the changes in muscle fiber and elasticfiber were not significant.

The angulated conformation of the aortic arch inyoung people is an anatomic difficulty that has onlyrecently received any attention. A new generation of endo-prosthesis is under clinical study and might help to over-come this problem. For example, the conformable-TAGendoprosthesis (W. L. Gore, Flagstaff, Ariz) is designedto adapt to severe aortic angulations. A clinical study ofthe conformable-TAG in trauma patients was started inNovember 2009.29 Cook Medical recently introducedthe Pro-Form release system (Cook Medical, Blooming-ton, Ind) in conjunction with low-profile endografts(TX2 LP) that can adapt to angulated arches with smallerdiameters. The Valiant endoprosthesis (MedtronicVascular) used in this study is being used in a clinical trialin trauma patients together with a new release system(Captivia; Medtronic Vascular), which can improve thestability and adaptation of the prosthesis to the conforma-tion of the aortic arch. The final data collection date for theprimary outcome (all patients are alive at 30 days afterimplantation) was scheduled for April 2012.3,30

Some authors have reported that incorporation of stentgrafts into the aortic wall is insufficient.31,32 Adequate

incorporation of the prosthesis is important to ensuregood tissue integration and to prevent migration. Thesestudies involved autopsy cases and surgery to explant anendoprosthesis in cases with aneurysms. Currently, thereis insufficient data to determine whether the endoprosthe-sis is fully incorporated in human aortas without atheroscle-rosis or aneurysm. Experimental studies in pigs revealedthat neointima formation in the aortic wall was evidentwithin 4 weeks and complete integration of the prosthesisoccurred within 12 weeks.19 We found evidence of a pseu-dointima covering the entire lumen of the stent, withcomplete incorporation of almost the entire endoprosthesisby 30 days (Fig 10, online only). This layer consists of anarray of muscle cells, collagen, and fibrin, similar to theneointima. We also identified mural thrombi formingbetween the prosthesis and the aortic wall. These changesare likely due to repair of the vascular wall in response tothe pressure exerted by self-expanding stents.25

Limitations. Several limitations of the study warrantmention. Although we used an experimental animal model(pigs) to avoid confounding factors, such as race, size,weight, and feeding, the biological characteristics of theaorta differ between pigs and humans. For example, pigsshow rapid aortic intima formation and extremely rapidincorporation of the prosthesis into the thoracic aorta.On the other hand, the aorta in this experimental modelis similar to that in young adults, as it lacks atherosclerosisand the diameters are much smaller than those of aneu-rysmal aortas. We only use one type of stent graft (Valiant;Medtronic Vascular), which has sufficient radial force toachieve a good anchorage and showed good wall apposi-tion close to an angulated aorta.14

Another variable is the duration of the study, whichrequired euthanasia of animals after 30 days to avoid rapidweight gain and limit aortic growth, which could influencethe results. Additionally, we only included 25 animals,which may limit our ability to extrapolate the data to actualclinical conditions.

CONCLUSIONS

The present histopathologic results confirmed ourpreviously reported results of biomechanical tests, whichshowed a tendency for a reduction in the force supportedby the aortic wall with increasing endograft oversizingand a loss of elasticity of the aortic wall with any degree

JOURNAL OF VASCULAR SURGERYVolume 58, Number 6 Sincos et al 1651

of oversizing. The histologic and immunohistochemicalstudies revealed that implantation of an endoprosthesiscaused structural disarrangement of the aortic wall, partic-ularly decreased amounts of muscle and elastic fibers.

AUTHOR CONTRIBUTIONS

Conception and design: IR, ES, MH, RAAnalysis and interpretation: IR, ES, SB, AB, MH, RAData collection: IR, ES, SB, ABWriting the article: IR, ES, AB, MH, RACritical revision of the article: IR, ES, AB, MH, VG,

JP, RAFinal approval of the article: IR, ES, SB, AB, MH, VG,

JP, RAStatistical analysis: IRObtained funding: IR, ES, JP, RAOverall responsibility: IR, ES, RA

REFERENCES

1. Clancy TV, Gary Maxwell J, Covington DL, Brinker CC, Blackman D.A statewide analysis of level I and II trauma centers for patients withmajor injuries. J Trauma 2001;51:346-51.

2. Sincos IR, Aun R, da Silva ES, Belczak S, de Lourdes Higuchi M,Gornati VC, et al. Impact of stent graft oversizing on the thoracic aorta:experimental study in a porcine model. J Endovasc Ther 2011;18:576-84.

3. Lee WA, Matsumura JS, Mitchell RS, Farber MA, Greenberg RK,Azizzadeh A, et al. Endovascular repair of traumatic thoracic aorticinjury: clinical practice guidelines of the Society for Vascular Surgery.J Vasc Surg 2011;53:187-92.

4. Mitchell ME, Rushton FW, Boland AB, Byrd TC, Baldwin ZK.Emergency procedures on the descending thoracic aorta in the endo-vascular era. J Vasc Surg 2011;54:1298-302; discussion: 1302.

5. Cannon RM, Trivedi JR, Pagni S, Dwivedi A, Bland JN, Slaughter MS,et al. Open repair of blunt thoracic aortic injury remains relevant in theendovascular era. J Am Coll Surg 2012;214:943-9.

6. Neschis DG, Scalea TM, Flinn WR, Griffith BP. Blunt aortic injury.N Engl J Med 2008;359:1708-16.

7. Sincos IR, Aun R, Belczak SQ, Nascimento LD, Mioto Netto B,Casella I, et al. Endovascular and open repair for blunt aortic injury,treated in one clinical institution in Brazil: a case series. Clinics (SaoPaulo) 2011;66:267-74.

8. Belczak S, Silva ES, Aun R, Sincos IR, Belon AR, Casella IB, et al.Endovascular treatment of peripheral arterial injury with covered stents:an experimental study in pigs. Clinics (São Paulo) 2011;66:1425-30.

9. Demetriades D, Velmahos GC, Scalea TM, Jurkovich GJ, Karmy-Jones R, Teixeira PG, et al. Operative repair or endovascular stent graftin blunt traumatic thoracic aortic inju-ries: results of an AmericanAssociation for the Surgery of Trauma Multi-center Study. J Trauma2008;64:561-70; discussion: 570-1.

10. Tang GL, Tehrani HY, Usman A, Katariya K, Otero C, Perez E, et al.Reduced mortality, paraplegia, and stroke with stent graft repair ofblunt aortic transections: a modern meta-analysis. J Vasc Surg 2008;47:671-5.

11. Idu MM, Reekers JA, Balm R, Ponsen KJ, de Mol BA, Legemate DA.Collapse of a stent graft following treatment of a traumatic thoracicaortic rupture. J Endovasc Ther 2005;12:503-7.

12. Muhs BE, Balm R, White GH, Verhagen HJ. Anatomic factors asso-ciated with acute endograft collapse after Gore TAG treatment ofthoracic aortic dissection or traumatic rupture. J Vasc Surg 2007;45:655-61.

13. Canaud L, Joyeux F, Berthet J-P, Hireche K, Marty-Ané C, Alric P.Impact of stent graft development on outcome of endovascular repairof acute traumatic transection of the thoracic aorta. J Endovasc Ther2011;18:485-90.

14. Canaud L, Alric P, Laurent M, Baum TP, Branchereau P, Marty-Ané CH, et al. Proximal fixation of thoracic stent grafts as a function ofoversizing and increasing aortic arch angulation in human cadavericaortas. J Endovasc Ther 2008;15:326-34.

15. Demetriades D, Velmahos GC, Scalea TM, Jurkovich GJ, Karmy-Jones R, Teixeira PG, et al. Diagnosis and treatment of blunt thoracicaortic injuries: changing perspectives. J Trauma 2008;64:1415-8.

16. Xenos ES, Abedi NN, Davenport DL, Minion DJ, Hamdallah O,Sorial EE, et al. Meta-analysis of endovascular vs open repair for trau-matic descending thoracic aortic rupture. J Vasc Surg 2008;48:1343-51.

17. Bandorski D, Brück M, Günther H-U, Manke C. Endograft collapseafter endovascular treatment for thoracic aortic disease. CardiovascIntervent Radiol 2010;33:492-7.

18. Svensson LG, Kouchoukos NT, Miller DC, Bavaria JE, Coselli JS,Curi MA, et al. Expert consensus document on the treatment ofdescending thoracic aortic disease. Ann Thorac Surg 2008;85(1 Suppl):S1-41.

19. Siegenthaler MP, Celik R, Haberstroh J, Bajona P, Goebel H,Brehm K, et al. Thoracic endovascular stent grafting inhibits aorticgrowth: an experimental study. Eur J Cardio-thorac Surg 2008;34:17-24.

20. Reis MM, Higuchi Mde L, Aiello VD, Benvenuti LA, Reis MM,Higuchi Mde L, et al. Growth factors in the myocardium of patientswith chronic chagasic cardiomyopathy. Rev Soc Bras Med Trop2000;33:509-18.

21. Koike MK, de Carvalho Frimm C, de Lourdes Higuchi M. BradykininB2 receptor antagonism attenuates inflammation, mast cell infiltrationand fibrosis in remote myocardium after infarction in rats. Clin ExpPharmacol Physiol 2005;32:1131-6.

22. Metzger M, Higuchi ML, Moreira LF, Chaves MJ, Castelli JB,Silvestre JM, et al. Relevance of apoptosis and cell proliferation forsurvival of patients with dilated cardiomyopathy undergoing partial leftventriculectomy. Eur J Clin Invest 2002;32:394-9.

23. Higuchi Mde L, Reis MM, Sambiase NV, Palomino SA, Castelli JB,Gutierrez PS, et al. Coinfection with Mycoplasma pneumoniae andChlamydia pneumoniae in ruptured plaques associated with acutemyocardial infarction. Arq Bras Cardiol 2003;81:12-22; 1-11.

24. Mohan D, Melvin JW. Failure properties of passive human aortic tissue.I-Uniaxial tension tests. J Biomech 1982;15:887-902.

25. White JG, Mulligan NJ, Gorin DR, D’Agostino R, Yucel EK,Menzoian JO. Response of normal aorta to endovascular grafting: a serialhistopathological study. Arch Surg (Chicago, Ill) 1998;133:246-9.

26. Forte A, Corte Della A, De Feo M, Cerasuolo F, Cipollaro M. Role ofmyofibroblasts in vascular remodelling: focus on restenosis and aneu-rysm. Cardiovasc Res 2010;88:395-405.

27. Scheumann J, Heilmann C, Beyersdorf F, Siepe M, Brenner RM,Böckler D, et al. Early histological changes in the porcine aortic mediaafter thoracic stent graft implantation. J Endovasc Ther 2012;19:363-9.

28. Wolinsky H, Glagov S. Nature of species differences in the medialdistribution of aortic vasa vasorum in mammals. Circ Res 1967;20:409-21.

29. Available at: http://clinicaltrials.gov/ct2/show/study/NCT00917852?term¼goreþtransections&rank¼1

30. Available at: http://clinicaltrials.gov/ct2/show/study/NCT01092767?term¼rescueþmedtronic&rank¼1

31. Shin CK, Rodino W, Kirwin JD, Ramirez JA, Wisselink W,Papierman G, et al. Histology and electron microscopy of explantedbifurcated endovascular aortic grafts: evidence of early incorporationand healing. J Endovasc Surg 1999;6:246-50.

32. Major A, Guidoin R, Soulez G, Gaboury LA, Cloutier G, Sapoval M,et al. Implant degradation and poor healing after endovascular repair ofabdominal aortic aneurysms: an analysis of explanted stent grafts.J Endovasc Ther 2006;13:457-67.

Submitted Oct 31, 2012; accepted Feb 1, 2013.

Additional material for this article may be found onlineat www.jvascsurg.org

Fig 5, online only. Van Gieson-stained tissue sections of an aorta in the control group vs group D. Black, elastic fiber.A, Cross-section of the aortic wall in control group (magnification, �4). B, Higher magnification view of the middlethird of the aortic wall in control group (magnification, �8). C, Cross-section of the aortic wall in group D(magnification, �2). D, Higher magnification view of the middle third of the aortic wall in group D(magnification, �20).

JOURNAL OF VASCULAR SURGERY1651.e1 Sincos et al December 2013

MEAN + C.I.95%CT A* B* C* D*

GROUP

10

20

30

40

50

60

70

ELAS

TIC

FIB

ER (

% )

Fig 6, online only. Comparison of relative elastic fiber content inthe aortic wall among the five groups (the percentage of field ofview stained black in uncompressed areas)‒Van Gieson-stained.Values are presented as the mean 6 95% confidence interval,*P < .001 vs control.

Fig 7, online only. Tissue sections of an aorta from group Dshowing evidence of an aneurysm. Black, elastic fiber. A, Cross-section of the aortic wall (magnification, �5; Masson). B,Higher magnification of the inner third of the aortic wall(magnification, �11; Masson). C, Van Gieson staining of thesection shown in (B).

JOURNAL OF VASCULAR SURGERYVolume 58, Number 6 Sincos et al 1651.e2

Fig 8, online only. a-actin-stained tissue sections of an aorta in the control group vs. group C. Yellow, smooth musclefibers in the endothelium. A, Cross-section of the aortic wall in control group (magnification, �4). B, Highermagnification view of the middle third of the aortic wall in control group (magnification, �8.6). C, Cross-section of theaortic wall showing the loss of a-actin in the inner layer in group C (magnification, �5). D, Higher magnification viewof the middle third of the aortic wall in group C (magnification, �11).

JOURNAL OF VASCULAR SURGERY1651.e3 Sincos et al December 2013

Fig 9, online only. TUNEL analysis of cross-sections of theaortic wall. A, Numerous apoptotic cells with green fluorescencecan be seen in group B (magnification, �4). B, No apoptotic cellscan be seen in a cross-section from the control group(magnification, �10).

Fig 10, online only. Incorporation of the stent graft and neo-intimal formation. A, Proximal view showing integration of theendoprosthesis around most of the aortic circumference. B,Proximal view showing partial adherence of the endoprosthesis tothe aortic circumference. C, Distal view showing neointimaformation on the stent and inner lumen of the aorta.

JOURNAL OF VASCULAR SURGERYVolume 58, Number 6 Sincos et al 1651.e4