Differential Tensile Strength and CollagenComposition in Ascending Aortic Aneurysmsby Aortic Valve PhenotypeJoseph E. Pichamuthu, MS,* Julie A. Phillippi, PhD,* Deborah A. Cleary, BS,Douglas W. Chew, BS, John Hempel, PhD, David A. Vorp, PhD, andThomas G. Gleason, MDDepartments of Bioengineering, Cardiothoracic Surgery, Surgery, Division of Vascular Surgery, and Biological Sciences,Center for Vascular Remodeling and Regeneration, and McGowan Institute for Regenerative Medicine, University ofPittsburgh, Pittsburgh, Pennsylvania

Background. Ascending thoracic aortic aneurysm(ATAA) predisposes patients to aortic dissection andhas been associated with diminished tensile strengthand disruption of collagen. Ascending thoracic aorticaneurysms arising in patients with bicuspid aortic valve(BAV) develop earlier than in those with tricuspid aorticvalves (TAV) and have a different risk of dissection.The purpose of this study was to compare aortic walltensile strength between BAV and TAV ATAAs anddetermine whether the collagen content of the ATAAwall is associated with tensile strength and valvephenotype.

Methods. Longitudinally and circumferentially ori-ented strips of ATAA tissue obtained during electivesurgery were stretched to failure, and collagen contentwas estimated by hydroxyproline assay. Experimentalstress-strain data were analyzed for failure strength andelastic mechanical variables: a, b, and maximal tangentialstiffness.

Accepted for publication July 1, 2013.

*These authors contributed equally to this work.

Address correspondence to Dr Gleason, C-700 Presbyterian University Hos-pital, 200 Lothrop St, Pittsburgh, PA 15213; e-mail: gleasontg@upmc.edu.

� 2013 by The Society of Thoracic SurgeonsPublished by Elsevier Inc

Results. The circumferential and longitudinal tensilestrengths were higher for BAV ATAAs when comparedwith TAV ATAAs. The a and b were lower for BAVATAAs when compared with TAV ATAAs. The maximaltangential stiffness was higher for circumferential whencompared with longitudinal orientation in both BAV andTAV ATAAs. The amount of hydroxyproline wasequivalent in BAV and TAV ATAA specimens. Althoughthere was a moderate correlation between the collagencontent and tensile strength for TAV, this correlation isnot present in BAV.Conclusions. The increased tensile strength and

decreased values of a and b in BAV ATAAs despiteuniform collagen content between groups indicate thatmicrostructural changes in collagen contribute to BAV-associated aortopathy.

(Ann Thorac Surg 2013;-:-–-)� 2013 by The Society of Thoracic Surgeons

icuspid aortic valve (BAV) is the most common

Bcongenital heart malformation occurring in 1% to 2%of the population and is associated with ascendingthoracic aortic aneurysm (ATAA) formation and asignificantly greater risk of aortic valve disease thanthe general population [1, 2]. Ascending thoracic aorticaneurysms progress to aortic dissection, rupture, andsudden death. Current expert surgical consensus con-cludes that patients with ATAA should undergo pro-phylactic surgery when the maximal orthogonal diameteris 48 to 55 mm, depending on several risk factors [3]. Incertain high-risk populations, such as patients with BAV,ATAAs have been known to dissect at smaller diameters[4], suggesting that diameter may not be the best pre-dictor of aortic catastrophe. Alternatively, because aortic

dissection or disruption is a biomechanical phenomenon,mechanical models should be developed to better predictrisk of aortic catastrophe.Although the mechanical properties of abdominal

aortic aneurysm have been widely studied, data forATAA are significantly more limited. Okamoto andcolleagues [5] demonstrated nonlinear behavior of ATAAunder biaxial testing and a decrease in distensibilitywith age. Our laboratory previously reported directionalstrength differences of ATAA compared with normalnonaneurysmal aorta [6]. Iliopoulos and associates [7]reported both regional and directional variations ofstrength and stiffness in ATAAs. Choudhury and co-workers [8] reported local stiffness changes from biaxialtests for ATAAs. To our knowledge, no information onaortic wall mechanical strength of ATAAs with respect tovalve phenotype has been reported.Mechanical behavior of the aortic wall depends on

extracellular matrix (ECM) structure and composition.Elastin and collagen determine the mechanical proper-ties of the aortic wall, with distensibility primarilydepending on elastin and peak aortic stiffness and tensile

0003-4975/$36.00http://dx.doi.org/10.1016/j.athoracsur.2013.07.001

Abbreviations and Acronyms

ATAA = ascending thoracic aortic aneurysmBAV = bicuspid aortic valveCIRC = circumferentialECM = extracellular matrixHYP = hydroxyprolineLONG = longitudinalMTS = maximal tangential stiffnessTAV = tricuspid aortic valve

2 PICHAMUTHU ET AL Ann Thorac SurgTENSILE STRENGTH IN ASCENDING AORTIC ANEURYSM 2013;-:-–-

strength provided by collagen [9]. The role of collagencontent in ATAA development remains unclear; someinvestigators have reported no differences in the elastinand collagen content of aneurysmal wall tissuesbetween patients with BAV and patients with tricuspidaortic valves (TAV) [10, 11], whereas others haveobserved regionally reduced collagen expression [12].No previous studies examined the relationship betweenthe ECM and the mechanical properties of ATAAs,or whether there is difference caused by aortic valvemorphology.

The purpose of this study was (1) to determine whetherthere is a difference in wall tensile strength between BAVand TAV ATAAs and (2) to determine the relativecollagen content within the ATAA wall and its relation-ship to mechanical properties of the two aortic valvephenotypes.

Material and Methods

Tissue HarvestWhole, fresh, nondissecting ascending aortic specimenswere harvested as an intact tubular structure (Fig 1A)from patients with BAV (age, 54 � 4 years; diameter, 50 �5 mm, mean � standard deviation; n ¼ 23) and TAV (age,66 � 11 years; diameter, 57 � 14 mm; n ¼ 15) undergoingelective surgery. All tissues were excised after obtaininginformed patient consent in accordance with a studyprotocol approved by our institutional review board.The tissue was harvested from above the sinuses of Val-salva with proximal and distal orientation noted by thesurgeon, immersed in saline solution, and kept refriger-ated at 4�C until testing. Patient demographics wereobtained from clinical records. The maximal diameter of

Fig 1. (A) Representative image of tubularascending thoracic aortic aneurysm tissueharvested. Suture indicates proximal end, anda tentative aorta axis is drawn in white dottedline. (B) Tubular tissue is cut open, andrectangular pieces of circumferentially(CIRC) and longitudinally (LONG) orientedspecimens are harvested for tensile testing.

each ATAA was determined y means of computedtomography.

Tensile TestingAscending thoracic aortic aneurysm tissues were testedwithin 48 hours of harvest. The ATAAs were preparedby first removing the adipose tissues and then dissectingcircumferentially (CIRC) and longitudinally (LONG)oriented strip specimens (Fig 1B). Specimen widthand thickness were measured at three different locationsusing a caliper (Scienceware, Wayne, NJ) and thenaveraged. Both ends of each strip were sandwiched be-tween the clamps using sandpaper and cyanoacrylateglue. The clamps were then mounted in a custom-builtuniaxial tensile testing machine [13] with a 25-poundload cell and amplifier (Transducer Techniques, Teme-cula, CA), and LabVIEW 7.1 (National Instruments,Austin, TX) was used to record measurements. Thespecimen was slowly stretched until a small increase inload was observed, and the initial specimen length wasnoted. The specimens were preconditioned and thenstretched until specimen failure as described previously[13]. The applied force and corresponding displacementwere collected synchronously and continuously at thesampling rate of 10 Hz until specimen failure.

Biomechanical Data AnalysisAs described previously [13], stress and strain werecalculated from force-displacement data and initialspecimen dimensions. Briefly, the Cauchy stress (T)within the specimen was calculated as the applied forcenormalized by deformed cross-sectional area, and thestrain (ε) was calculated as deformation normalized byoriginal specimen length. The maximal tangential stiff-ness (MTS) was taken as the maximal slope (ie, maximalresistance to deformation), and the tensile strength wastaken as the peak stress obtained before failure from thestress-strain curve [6].We used our prior model to further characterize the

biomechanical response of the aneurysm wall [14]. Themodel relates the stress in a uniaxial loaded specimen tothe stretch (l ¼ ε þ 1) through the following equation:

T ¼h2aþ 4b

�l2 þ 2l�1 � 3

�ihl2 � l�1

i(1)

where a and b are model parameters indicative of themechanical properties of the tissue. The mathematical

3Ann Thorac Surg PICHAMUTHU ET AL2013;-:-–- TENSILE STRENGTH IN ASCENDING AORTIC ANEURYSM

model was fit to the elastic region of each experimentallyderived T – l data sets (ie, before yield) using theMarquardt-Levenberg least squares nonlinear regressionalgorithm in Sigma Plot 11.0 (Systat Software, Inc, SanJose, CA). A representative model was fit for stress-stretch data (Fig 2).

Collagen Content AssessmentTRICHROME STAINING. Aortic specimens (approximately 1 �0.5 cm2) were fixed in 10% buffered formalin andembedded in paraffin. Four-micrometer sections werestained using Masson’s trichrome (Research HistologyServices, University of Pittsburgh Thomas E. StarzlTransplantation Institute, Pittsburgh, PA) to qualitativelyassess collagen composition. Slides were visualized usinga Nikon TE-2000-E inverted microscope and capturedusing a Nikon DS-Fi1 5MP color camera and NIS Ele-ments Software (Nikon, Melville, NY).HYDROXYPROLINE. Total collagen content was determinedfor a subset of samples from each group using a modifi-cation of a hydroxyproline (HYP) assay [15, 16]. Aftertensile testing, a small portion of flanking tissue (50 to 100mg) from both sides of the failure region was minced(approximately 1 � 1 mm) specimens (n ¼ 49 for BAV;n ¼ 26 for TAV) and lyophilized. Specimens were flame-sealed in 10-mm Pyrex tubes with 100 mL of 6N HClcontaining 0.5% (v/v) phenol, incubated at 110�C for24 hours, and dried under vacuum. The dried sample wasresuspended with 1% (v/v) HCl, incubated at 50�C, andthen neutralized with an equivalent amount of NaOH.Chloramine T reagent (0.056 mol/L) was added, followedby incubation at room temperature for 25 minutes.Freshly prepared Ehrlich’s reagent (1 mol/L) was added,followed by incubation at 65�C for 20 minutes. Absor-bance of replicates was measured at 550 nm in aspectrophotometer (SpectraMax M2, model D02667, Mo-lecular Devices, LLC, Sunnyvale, CA). Experimental HYPlevels were normalized to the mass of dry tissue.

Fig 2. A representative fit of the model given by Equation 1 to theexperimental stress-stretch data obtained from the tensile testing of anascending thoracic aortic aneurysm tissue specimen. This particulardata is from a longitudinally oriented specimen of an ascendingthoracic aortic aneurysm with bicuspid aortic valve.

Statistical AnalysisAll data are presented as mean � standard error of themean. Student’s t test was performed to compare themean values between the groups using Sigma Plot 11.0(Systat Software Inc, Chicago, IL). Significance wasassumed for a probability value of less than 0.05.

Results

A total of 23 ATAAs with BAV and 15 ATAAs with TAVwere included in this study. The average maximalorthogonal diameter of the aorta was 50 � 5 mm for BAVversus 57 � 14 mm for TAV (p ¼ 0.15). The age discrep-ancy noted is consistent with the clinical observation thatBAV ATAAs present 10 to 20 years earlier than patientswith TAV and ATAA [17]. From these ATAAs, 178 sam-ples were tensile tested; 15 of those were either slipped orbroke at the clamp, and hence were discarded from dataanalysis. There was no significant difference between thetensile test data normalized per patient and the non-normalized data of all specimens averaged for the studygroup. Thus, nonnormalized values averaged for thewhole group are reported here.For BAV tissues, the CIRC and LONG tensile strengths

were 165.6 � 9.8 N/cm2 and 69.8 � 3.1 N/cm2, respectively(Table 1; Fig 3). For TAV, these were 96.1 � 6.1 N/cm2 and54.0 � 3.7 N/cm2, respectively. In both BAV and TAV, theCIRC oriented samples were stronger than LONG ori-ented samples (p < 0.05), and the strength of BAV wasstronger than TAV in both orientations (p < 0.05). Theaverage percentage of failure strain for BAV in CIRC andLONG orientations is shown in Table 1.The MTS of BAV appeared to be higher when

compared with TAV CIRC specimens and lower thanTAV in LONG specimens (Table 1). In BAV the MTS was350.4 � 16.0 N/cm2 and 191.6 � 9.6 N/cm2 for CIRC andLONG specimens (p < 0.05), respectively. Similarly, inTAV the MTS was 335.1 � 22.2 N/cm2 and 220.7 � 20.3N/cm2 for CIRC and LONG specimens (p < 0.05),respectively.The average closeness-of-fit of the biomechanical model

described in Equation 1 to the experimental stress-stretchdata was found to be greater than 99% (r2 > 0.99 � 0.001).The mean value of a was not significantly differentin LONG and CIRC specimens for either BAV or TAV(Table 1; TAV, 6.7 � 2.3 N/cm2 versus 6.2 � 0.8 N/cm2;BAV, 4.8� 0.3N/cm2 versus 4.5� 0.4N/cm2).However,a issignificantly lower in BAV LONG specimens comparedwith TAV LONG (p < 0.05). The b is not significantlydifferent between CIRC and LONG specimens in TAV(Table 1; 71.0 � 22.2 N/cm2 versus 120.1 � 31.5 N/cm2; p ¼0.1), but was significantly different within BAV (12.3 � 1.1N/cm2 versus 18.1� 1.8N/cm2; p< 0.005). In addition, the bis significantly lower in BAV compared with TAV in bothorientations (p < 0.005).Trichrome staining of ATAAs revealed cystic media

degeneration, elastin fragmentation, and a dispropor-tionate distribution of collagen among BAV and TAV,suggesting increased collagen in BAV (Fig 4). However,

Table 1. Biomechanical Properties and Hydroxyproline Values Assessed for Bicuspid Aortic Valve And Tricuspid Aortic Valve

Variable a b MTS Failure Strain (%) Tensile Strength HYP

TAV CIRC 6.8 � 2.3 71.0 � 22.2 335.1 � 22.2a 60.7 � 4.1a 96.1 � 6.1a 19.0 � 1.5 (all TAV)TAV LONG 6.2 � 0.8b 120.0 � 31.5b 220.7 � 20.3 46.9 � 3.1b 54.0 � 3.7b

BAV CIRC 4.8 � 0.3 12.3 � 1.1c 350.4 � 16.0 91.7 � 3.7c 165.6 � 9.8c 18.9 � 1.9 (all BAV)BAV LONG 4.5 � 0.4 18.1 � 1.8d 191.6 � 9.6d 63.4 � 2.2d 69.8 � 3.1d

a p < 0.05 TAV CIRC compared with TAV LONG. b p < 0.05 TAV LONG compared with BAV LONG. c p < 0.05 BAV CIRC compared with TAVCIRC. d p < 0.05 BAV LONG compared with BAV CIRC.

BAV ¼ bicuspid aortic valve; CIRC ¼ circumferential; HYP ¼ hydroxyproline; LONG ¼ longitudinal; MTS ¼ maximal tangentialstiffness; TAV ¼ tricuspid aortic valve.

4 PICHAMUTHU ET AL Ann Thorac SurgTENSILE STRENGTH IN ASCENDING AORTIC ANEURYSM 2013;-:-–-

there was no difference in HYP for BAV and TAV(18.9 mg/mg � 1.9 versus 19.0 � 1.5, respectively; p ¼ 0.17;Table 1). For TAV, a moderate correlation between thecollagen content and tissue strength was demonstrated(r ¼ 0.359; Fig 5), but this correlation was not seen in BAVtissues (r ¼ 0.014).

Comment

Aortic dissection or rupture represents a mechanical fail-ure of the aortic wall, which can occur when aorticwall integrity is altered. Despite a higher percentage ofpatients with BAV compared with TAV among the aorticdissection population than among the general population[18], no investigation has been done, to our knowledge,on the condition of ATAA wall tensile strength withrespect to valve phenotype. This study found differences inthe wall strength between ATAAs from BAV and TAVpatients. The tensile strength in the CIRC orientation washigher than in the LONG orientation for both BAV andTAV ATAAs. Tensile strength of BAV ATAAs was higherwhen compared with that of TAV ATAAs for boththe CIRC and LONG orientations. The directional

Fig 3. The average tensile strength ofbicuspid aortic valve (BAV) and tricuspidaortic valve (TAV) in circumferential (CIRC)and longitudinal (LONG) orientations. UThestrength of circumferentially orientedspecimen is significantly higher thanlongitudinally oriented specimen for bothbicuspid aortic valve (p < 0.01) and tricuspidaortic valve (p < 0.01). **Bicuspid aorticvalve is significantly stronger than tricuspidaortic valve in both orientations (p < 0.01).

differences in the tensile strength of the ATAA wall areconsistent with the demonstrated anisotropic nature of theATAA wall as previously reported by our group andothers and is consistent with lower tensile strength ofabdominal aortic aneurysms (AAAs) and ATAAs whencompared with normal aortas [5–7, 19, 20]. Moreover,Iliopoulos and associates [7] reported spatial variation inthe material variables such as strength and stiffness ofATAAs.Our qualitative, histologic observation of increased

collagen in BAV compared with TAV (Fig 4) agrees with aprior report from our team that demonstrated elevatedtype I collagen mRNA [21], and from data reported byChoudhury and colleagues [8]. Further, Wang and co-workers [22] demonstrated increased collagen depositionin aortic dissection. Iliopoulos and colleagues [19] docu-mented higher levels of collagen histologically in ATAAswhen compared with normal aortas. On the other hand,our quantitative assessment of collagen content by HYPassay showed no difference between BAV and TAV,which is consistent with reports by Fedak and associates[10] and LeMaire and coworkers [11]. Collectively,these data suggest that differences in mechanical

Fig 4. Three representative micrographs ofMasson’s trichrome staining of bicuspidaortic valve (A, C, E) and tricuspid aorticvalve (B, D, F) tissues. Note that bicuspidaortic valve shows more disrupted collagen(blue color, white arrows). Tricuspid aorticvalve has less collagen (blue color, blackarrows), but aligned in orientation. Scalebar ¼ 50 mm.

5Ann Thorac Surg PICHAMUTHU ET AL2013;-:-–- TENSILE STRENGTH IN ASCENDING AORTIC ANEURYSM

properties are not attributable to absolute collagen con-tent, but may be accounted for by microstructuralchanges in the collagen framework within the ECM (ie,collagen architecture) as we recently revealed in nativeaortic specimens [23].

Although the tensile strength of BAV is significantlystronger when compared with TAV in both CIRC andLONG orientations, the total HYP was equivalent in bothgroups. The finding of differential correlation betweencollagen content and tensile strength in TAV (r ¼ 0.359)compared with BAV (r ¼ 0.014) may indicate distin-guishable architecture. Our finding suggests the propor-tional differences in tensile strength between BAV andTAV cannot be explained by alterations in total collagencontent.

The aortic walls of BAV and TAV exhibit differencesin mechanical behavior as characterized by ournonlinear elastic biomechanical model. The materialvariables (a, b) represent the initial stiffness of thematerial in the nonlinear stress-stretch curve in theelastic region up to MTS for each specimen tested: arepresents the shear modulus for infinitesimal defor-mation of structural fibers, and b is a stiffening variable

or bulk modulus of the tissue. We tested this model topredict which of these two variables influences thestress-stretch curves with one of the variables keptconstant while varying the other variable. We foundthat b influences the model 12 times more than a whenthe difference in simulated stress for incremental bincrease was divided by the difference in stress for thecorresponding incremental a increase at the maximalfailure stretch of 1.9. For BAV, a is lower by 28% and26%, in the CIRC and LONG orientations, respectively,and b is lower by 83% and 85%, respectively, whencompared with TAV. The fact that a and b weremuch lower with BAV when compared with TAV in-dicates that fewer structural fibers were engaged tobear the initial stretch load in BAV. This low a and b inBAV suggests that collagen may be more coiled ordisrupted as a result of microstructural changes of thecollagen network within the ECM. b is higher in LONGwhen compared with CIRC for BAV and TAV. Asignificantly lower value of b in CIRC for BAV indicatesa possibility of immature organization of collagen fi-bers in CIRC. Furthermore, b within TAV betweenCIRC and LONG oriented specimens does not vary

Fig 5. Collagen content, as estimated by hydroxyproline (HYP), as afunction of tensile strength for each tricuspid aortic valve (TAV; A)and bicuspid aortic valve (BAV; B) specimen tested. Note a moderatecorrelation (r ¼ 0.359) between the collagen content and strength fortricuspid aortic valve, but not for bicuspid aortic valve (r ¼ 0.014).

6 PICHAMUTHU ET AL Ann Thorac SurgTENSILE STRENGTH IN ASCENDING AORTIC ANEURYSM 2013;-:-–-

significantly, suggesting a more mature organization ofcollagen fibers in both orientations for TAV, which isconsistent with our recent observation of increasedmature collagen in TAV versus BAV ATAAs [23].Okamoto and associates [5] reported differences in theproperties of BAV and TAV using an exponentialconstitutive model in biaxial testing, which wereconsistent with reduced a and b in BAV comparedwith TAV. Previously we reported no difference in aand b between LONG and CIRC of ascendingaortic aneurysms, but ATAAs showed a differencebetween LONG and CIRC for b [6, 13]. Both a and b ofATAAs were reduced when compared with ascendingaortic aneurysms, and for ATAAs b is higher in LONG(53 N/cm2) than in CIRC (17 N/cm2). A similar trend isnow observed in BAV and TAV as shown in Table 1.However, the value of b is lower in BAV and higher inTAV compared with the average for all ATAAs re-ported earlier in both LONG and CIRC oriented di-rection. The current data suggest that the ATAA wallmaterial property characteristics are strongly associ-ated with valve phenotype. Equivalent total collagencontent despite different biomechanical material vari-ables suggests there are distinct architectural

differences of the ECM in the two phenotypes thatimpart the mechanical property differences.Previously, we observed increased MTS in ATAAs

when compared with normal aortas [6]. In the currentstudy, we showed that the MTS measured in the failureregion is higher in CIRC than in LONG for both TAVand BAV, which seems contrary to the stiffnessmeasured in the elastic region in both cases. This may bebecause at higher stretch, all the structural fibers of thetissue may be engaged or aligned in the preferred di-rection to resist deformation. Circumferential MTS ofBAV was the highest and LONG MTS of BAV was thelowest when compared with the corresponding orien-tation of TAV. Higher MTS of CIRC indicates that BAVATAAs become stiffer after initial stretching in thecircumferential direction, and lower MTS of LONG in-dicates that BAV ATAAs become more compliant in thelongitudinal direction. The higher difference betweenCIRC and LONG stiffness of BAV implies that BAVATAAs are more anisotropic when compared with TAVATAAs. The increased anisotropic nature of BAVcould explain why BAV ATAAs tend to develop in thecharacteristically asymmetric pattern that they do alongthe greater curvature of the proximal ascending aorta.To date, our observations of aortic wall biomechanicsincluding alterations in tensile strength, elastic consti-tutive material modeling, and tangential stiffness allsupport our overarching hypothesis that aneurysm for-mation in BAV patients is caused by alterations in ECMarchitecture rather than collagen content.A potential limitation of the study is the age discrep-

ancy between patients with BAV and TAV ATAAs, suchthat aging may compound architectural changes [4, 17].However, we have previously shown that patient age andaneurysm diameter do not impact the delaminationstrength of the aortic media [24]. We proposed thatreduced medial delamination strength could contribute toa greater risk of dissection in BAV compared with TAVATAA patients [24]. In this study, we linearly extrapolatedthe strength of BAV to the corresponding age of TAV.When the tensile strength of BAV was extrapolated to thatof the age of TAV, strength of BAV remained higher thanTAV (Fig 6).This study is the first to our knowledge to compare

material property, directionally dependent strength,and collagen content in the same ATAA specimensbetween BAV and TAV patients. We showed that BAVATAAs have greater tensile strength in the CIRC andLONG directions and greater failure strain than TAVATAAs. These data suggest that absolute collagencontent does not explain the differences found in themechanical properties tested for BAV and TAV ATAAs.Rather, given the marked differences in both strengthand failure strain, these biomechanical differencessupport the notion that microstructural differencesarise in the collagen framework of the ECM in BAVcompared with TAV. Ongoing investigations arefocused on the role that matrix architecture plays inregion-specific biomechanical strength of the BAV aortaand its pathologic propensity.

Fig 6. The extrapolated bicuspid aortic valve(BAV) tensile strength for the varioustricuspid aortic valve (TAV) patient ages.*The extrapolated tensile strength of bicuspidaortic valve is significantly higher thantricuspid aortic valve in both orientations(circumferential [CIRC], p < 0.01; longitu-dinal [LONG], p < 0.01).

7Ann Thorac Surg PICHAMUTHU ET AL2013;-:-–- TENSILE STRENGTH IN ASCENDING AORTIC ANEURYSM

The authors thank Benjamin Green and Michael Eskay for theirgenerous help in tissue collection and transportation and KristinValchar, Michelle Shirey and Erica Butler for their assistance inIRB procedures and informed patient consent. This researchwork was supported by grants R01-HL060670-09 (D.A.V.), R01-HL086418-04 (D.A.V.) and R01-HL109132-01 (T.G.G.) from theNational Heart, Lung, and Blood Institute of the NationalInstitutes of Health. We certify that this manuscript has not beenpreviously published and is not being considered in whole or inpart in any other scientific journal. All authors are aware of andagree to the content of the manuscript. None of the authors haveany conflict of interest to declare.

References

1. Braverman AC, Guven H, Beardslee MA, Makan M,Kates AM, Moon MR. The bicuspid aortic valve. Curr ProblCardiol 2005;30:470–522.

2. Fedak PW, Verma S, David TE, Leask RL, Weisel RD,Butany J. Clinical and pathophysiological implications of abicuspid aortic valve. Circulation 2002;106:900–4.

3. Ergin MA, Spielvogel D, Apaydin A, et al. Surgical treatmentof the dilated ascending aorta: When and how? Ann ThoracSurg 1999;67:1834–9; discussion 1853–6.

4. Tadros TM, Klein MD, Shapira OM. Ascending aortic dila-tation associated with bicuspid aortic valve: Pathophysi-ology, molecular biology, and clinical implications.Circulation 2009;119:880–90.

5. Okamoto RJ, Wagenseil JE, DeLong WR, Peterson SJ,Kouchoukos NT, Sundt TM 3rd. Mechanical properties ofdilated human ascending aorta. Ann Biomed Eng 2002;30:624–35.

6. Vorp DA, Schiro BJ, Ehrlich MP, Juvonen TS, Ergin MA,Griffith BP. Effect of aneurysm on the tensile strength andbiomechanical behavior of the ascending thoracic aorta. AnnThorac Surg 2003;75:1210–4.

7. Iliopoulos DC, Deveja RP, Kritharis EP, et al. Regional anddirectional variations in themechanical properties of ascendingthoracic aortic aneurysms. Med Eng Phys 2009;31:1–9.

8. Choudhury N, Bouchot O, Rouleau L, et al. Local me-chanical and structural properties of healthy and diseasedhuman ascending aorta tissue. Cardiovasc Pathol 2009;18:83–91.

9. Sokolis DP, Kefaloyannis EM, Kouloukoussa M, Marinos E,Boudoulas H, Karayannacos PE. A structural basis for theaortic stress-strain relation in uniaxial tension. J Biomech2006;39:1651–62.

10. Fedak PW, de Sa MP, Verma S, et al. Vascular matrix remod-eling in patients with bicuspid aortic valve malformations:implications for aortic dilatation. J Thorac Cardiovasc Surg2003;126:797–806.

11. LeMaire SA, Wang X, Wilks JA, et al. Matrix metal-loproteinases in ascending aortic aneurysms: Bicuspid versustrileaflet aortic valves. J Surg Res 2005;123:40–8.

12. Cotrufo M, Della Corte A, De Santo LS, et al. Differentpatterns of extracellular matrix protein expression in theconvexity and the concavity of the dilated aorta with bicuspidaortic valve: preliminary results. J Thorac Cardiovasc Surg2005;130:504–11.

13. RaghavanML, Webster MW, Vorp DA. Ex vivo biomechanicalbehavior of abdominal aortic aneurysm: Assessment using anew mathematical model. Ann Biomed Eng 1996;24:573–82.

14. Raghavan ML, Vorp DA. Toward a biomechanical tool toevaluate rupture potential of abdominal aortic aneurysm:Identification of a finite strain constitutive model and eval-uation of its applicability. J Biomech 2000;33:475–82.

15. Stegemann H. Microdetermination of hydroxyproline withchloramine-t and p-dimethylaminobenzaldehyde [inGerman].Hoppe Seylers Z Physiol Chem 1958;311:41–5.

16. Stegemann H, Stalder K. Determination of hydroxyproline.Clin Chim Acta 1967;18:267–73.

17. Gleason TG. Heritable disorders predisposing to aorticdissection. Semin Thorac Cardiovasc Surg 2005;17:274–81.

18. Larson EW, Edwards WD. Risk factors for aortic dissection: anecropsy study of 161 cases. Am J Cardiol 1984;53:849–55.

19. Iliopoulos DC, Kritharis EP, Giagini AT, Papadodima SA,Sokolis DP. Ascending thoracic aortic aneurysms are asso-ciated with compositional remodeling and vessel stiffeningbut not weakening in age-matched subjects. J Thorac Car-diovasc Surg 2009;137:101–9.

20. Okamoto RJ, Xu H, Kouchoukos NT, Moon MR,Sundt TM 3rd. The influence of mechanical properties onwall stress and distensibility of the dilated ascending aorta.J Thorac Cardiovasc Surg 2003;126:842–50.

21. Phillippi JA, Eskay MA, Kubala AA, Pitt BR, Gleason TG.Altered oxidative stress responses and increased type icollagen expression in bicuspid aortic valve patients. AnnThorac Surg 2010;90:1893–8.

8 PICHAMUTHU ET AL Ann Thorac SurgTENSILE STRENGTH IN ASCENDING AORTIC ANEURYSM 2013;-:-–-

22. Wang X, LeMaire SA, Chen L, et al. Increased collagendeposition and elevated expression of connective tissuegrowth factor in human thoracic aortic dissection. Circula-tion 2006;114(1 Suppl):I200–5.

23. Phillippi JA, Green BR, Eskay MA, et al. Mechanism of aorticmedial matrix remodeling is distinct in bicuspid aortic valve

patients. J Thorac Cardiovasc Surg 2013 Jun 10 [Epub aheadof print].

24. Pasta S, Phillippi JA, Gleason TG, Vorp DA. Effect of aneu-rysm on the mechanical dissection properties of the humanascending thoracic aorta. J Thorac Cardiovasc Surg 2012;143:460–7.