PERIPHERAL ARTERIALINTERVENTIONSDiagnosis

Pedersen G, Laxdal E. Flow measure-ment before and after papaverine in-jection in above-knee prosthetic femo-ropopliteal bypass. J Vasc Surg 2006;43(4):729–734.

• OBJECTIVE: To investigate thevalue of intraoperative blood flow mea-surements on early and long-term pa-tency of above-knee prosthetic femoro-popliteal bypass. METHODS: Flow wasmeasured with a transit time flowmeterbefore (basal flow) and after an intra-graft injection of papaverine (papaverineflow) in 87 operations (86 patients) be-tween January 1990 and December 2001.Sixty-one grafts were of polyester, and26 were of polytetrafluoroethylene. Theoperations were done under epiduralanesthesia. The preoperative angio-graphic run-off score and clinical riskfactors were recorded. Patency rateswere analyzed with the product limitmethod and compared with the log-ranktest. Variables found to be near signifi-cantly related to patency rates (P � .1)were included in a multivariate analysisperformed with the Cox proportionalhazard model. RESULTS: Basal flowmeasurements were not related to pa-tency. The 2- and 5-year patency ratesfor grafts with a papaverine flow � 500mL/min were 48% and 18% comparedwith 66% and 52% for grafts with a pa-paverine flow � 500 mL/min. These dif-ferences were statistically significant (P� .012, hazard ratio, 2.6). Two- and5-year patency rates for smokers vs non-smokers were 44% and 18% vs 69% and54%. The patency rates for patients withpoor vs good run-off were 42% and 27%vs 66% and 31%. Smoking (P � .008,hazard ratio, 2.75) and poor run-off score(P � .009, hazard ratio, 2.38) were foundto be independent risk factors for re-duced patency rates. Poor run-off scoredid not correlate with low values of mea-sured basal or papaverine flow. CON-CLUSIONS: Papaverine flow of � 500mL/min is associated with reduced mid-and long-term patency rates. Additionalantithrombotic medication and frequentfollow-up for these grafts should be con-sidered. The inferior patency rates ofsmokers and patients with poor run-offindicate that prosthetic bypass is lesssuitable for these groups of patients.Authors’ Abstract

Musicant SE, Taylor LM, Jr. Prospec-tive evaluation of the relationship be-tween C-reactive protein, D-dimer andprogression of peripheral arterial dis-ease. J Vasc Surg 2006; 43(4):772–780;discussion 780.

• OBJECTIVE: Elevated levels of C-reactive protein (CRP) and D-dimer(DD) have been associated with the pres-ence and progression of various forms ofatherosclerotic disease, particularly cor-onary heart disease. We hypothesize thatthere is a relationship between elevatedlevels of baseline CRP and DD and pro-gression of peripheral arterial disease(PAD) in patients with symptomaticPAD. The current study is a prospectiveevaluation of this hypothesis. METH-ODS: Between 1996 and 2003, 384 sub-jects were enrolled in a National Insti-tutes of Health-sponsored blinded,prospective trial evaluating the effects ofmultiple atherosclerotic risk factors onprogression of symptomatic PAD. Base-line levels of CRP and D-dimer wereobtained in 332 subjects. Subjects werefollowed every 6 months with clinicalhistory and exam, ankle-brachial pres-sure index (ABI), and carotid artery du-plex scanning (CDS). The primary studyend point was a composite of ABI pro-gression, CDS progression, stroke, myo-cardial infarction, amputation, anddeath from cardiovascular disease. Sec-ondary end points included each of thecomponents of the primary end point.The relationship between time to thevarious endpoints and baseline CRP andDD levels was examined by life-tableanalysis and Cox proportional hazardsanalysis. RESULTS: Adequate baselinesamples for CRP and DD were availablein 332 subjects (mean age, 67 years;57.8% men) with mean follow-up of 38.4months (range, 1 to 99 months). Meanbaseline levels (� SD) for CRP were 0.8� 1.14 (range, 0.03 to 13.0), and meanDD levels were 227.4 � 303.3 (range, 1.9to 2744.8). Progression, as defined by theprimary end point, occurred in 48.5% ofsubjects. Subjects with elevated CRP(highest tertile) were no more likely tohave any of the progression end pointsthan those with the lowest values (low-est tertile) (P � NS, log-rank test, for allcomparisons). By univariate analysis,subjects with elevated DD (highest ter-tile) were significantly more likely to diefrom any cause compared with subjectswith the lowest DD values (lowest ter-tile) (P � .03, log-rank test). They were,

however, no more likely to reach any ofthe other progression end points, includ-ing the primary end point (P � NS, log-rank test for all other comparisons). Mul-tivariate analysis showed that DD levelwas a significant independent variableassociated with occurrence of myocar-dial infarction (hazard ratio, 2.3; P �.02). CONCLUSIONS: In subjects withsymptomatic PAD, elevated baselineDD, a marker of thrombotic activity, wassignificantly associated with the occur-rence of myocardial infarction. Thisstudy did not confirm a relationship be-tween progression of PAD and baselineDD or CRP during the first 3 years. Base-line DD and CRP do not provide usefulrisk stratification in patients at high riskfor progression of symptomatic PAD.Future studies should evaluate seriallevels of these markers to assess theirutility in predicting progression ofsymptomatic PAD.Authors’ Abstract

Angioplasty

Batt M, Baque J. Percutaneous angio-plasty of the superior gluteal arteryfor buttock claudication: a report ofseven cases and literature review. JVasc Surg 2006; 43(5):987–991.

• BACKGROUND: Buttock claudica-tion due to stenosis or occlusion of thesuperior gluteal artery is infrequent. Therecent development of noninvasive glu-teal duplex scanning, combined withaortoiliac angiography using obliqueprojections and the availability of low-profile devices for percutaneous translu-minal angioplasty (PTA), led us to re-view our recent experience concerningthe diagnosis and mid-term results ofPTA for superior gluteal artery stenosisor occlusion. METHODS: The files of allpatients who had been treated in ourdepartment by PTA for superior glutealartery stenosis or occlusion with buttockclaudication were analyzed retrospec-tively, and any associated arterial le-sions, morbidity, restenosis, or recurrentbuttock claudication were noted. Out-comes were compared with publishedreports. RESULTS: Retrospective reviewidentified six patients (5 men, 1 woman;mean age, 64 years) with seven cases ofbuttock claudication (1 bilateral localiza-tion) who had undergone PTA withinthe past 2 years. There was no case ofisolated buttock claudication. Buttockclaudication was associated with impo-

Abstracts of Current Literature

1216

tence, thigh claudication, or calf claudi-cation in seven cases. Gluteal duplexscans were performed for three of thepatients diagnosed with two stenosesand one occlusion. Aortoiliac angiogra-phy revealed five superior gluteal arterystenoses and two occlusions. PTA with-out stenting was successful in all cases,without morbidity or mortality. Duringa mean follow-up of 13 months, resteno-sis occurred in one patient. A repeat PTAwithout stenting was successful, withresolution of the buttock claudication.CONCLUSIONS: Buttock claudicationdue to superior gluteal artery stenosis isprobably underestimated when glutealduplex scanning and aortoiliac angiog-raphy with oblique projections are notperformed. PTA gives good results, andthe procedure can be repeated shouldrestenosis occur.Authors’ Abstract

Stents

Schillinger M, Sabeti S. Balloon an-gioplasty versus implantation of niti-nol stents in the superficial femoralartery. N Engl J Med 2006; 354(18):1879–1888.

• BACKGROUND: Because stent im-plantation for disease of the superficialfemoral artery has been associated withhigh rates of late clinical failure, percu-taneous transluminal angioplasty is pre-ferred for endovascular treatment, andstenting is recommended only in theevent of suboptimal technical results.We evaluated whether primary implan-tation of a self-expanding nitinol (nickel-titanium) stent yielded anatomical andclinical benefits superior to those af-forded by percutaneous transluminalangioplasty with optional secondarystenting. METHODS: We randomly as-signed 104 patients who had severe clau-dication or chronic limb ischemia due tostenosis or occlusion of the superficialfemoral artery to undergo primary stentimplantation (51 patients) or angioplasty(53 patients). Restenosis and clinical out-comes were assessed at 6 and 12 months.RESULTS: The mean (�SD) length of thetreated segment was 132 � 71 mm in thestent group and 127 � 55 mm in theangioplasty group. Secondary stentingwas performed in 17 of 53 patients (32percent) in the angioplasty group, inmost cases because of a suboptimal re-sult after angioplasty. At 6 months, therate of restenosis on angiography was 24percent in the stent group and 43 percentin the angioplasty group (P � 0.05); at 12months the rates on duplex ultrasonog-raphy were 37 percent and 63 percent,respectively (P � 0.01). Patients in thestent group were able to walk signifi-cantly farther on a treadmill at 6 and 12

months than those in the angioplastygroup. CONCLUSIONS: In the interme-diate term, treatment of superficial-fem-oral-artery disease by primary implanta-tion of a self-expanding nitinol stentyielded results that were superior tothose with the currently recommendedapproach of balloon angioplasty withoptional secondary stenting. (Clinical-Trials.gov number, NCT00281060.)Authors’ Abstract

Stent-grafts

Sivamurthy N, Schneider DB. Ad-junctive primary stenting of Zenithendograft limbs during endovascularabdominal aortic aneurysm repair: im-plications for limb patency. J VascSurg 2006; 43(4):662–670.

• OBJECTIVE: Endograft limb occlu-sion is an infrequent but serious compli-cation of endovascular abdominal aorticaneurysm (AAA) repair. The insertion ofadditional stents within the endograftlimb may prevent future occlusion. Thisstudy evaluates limb patency with andwithout adjunctive stenting of endograftlimbs at the time of endovascular AAArepair. METHODS: We performed a ret-rospective review of 248 patients whounderwent endovascular abdominalaortic aneurysm repair with the ZenithAAA endovascular graft between 1999and 2004. Among these patients, twogroups were identified: 64 patients withadjunctive stents placed in 85 limbs and184 patients without additional barestent placement in endograft limbs at thetime of endovascular AAA repair. RE-SULTS: Women comprised 23% ofstented and 11% of unstented patients (P� .02). The mean length of follow-up inthe stented and unstented groups was2.0 years. There were 13 instances oflimb thrombosis in 13 patients (5.2% ofpatients, 2.7% of limbs), all in the un-stented group. No limb occlusions oc-curred in the presence of adjunctive baremetal stents. Seventy-three percent ofthe occlusions occurred � 6 months ofendovascular AAA repair. Two patients(15%) had no symptoms of lower-extremity ischemia despite graft limb oc-clusion and did not undergo interven-tion. The others underwent thrombec-tomy (n � 2), thrombectomy with barestent placement (n � 3), femoral-femoralbypass (n � 4), thrombolysis (n � 1), andthrombolysis with bare stent placement(n � 1). Of the seven who underwentthrombectomy or thrombolysis, threehad no additional stents placed at thesecondary procedure, and two of thesethree went on to rethrombose. By life-table analysis, primary patency at 3years in the stented and nonstentedlimbs was 100% � 0% and 94% � 3%,

respectively (P � .05). CONCLUSIONS:The intraoperative insertion of addi-tional bare metal stents appeared toeliminate the risk of thrombosis and waswithout complication. Of the 85 stentedlimbs in this series, not one occluded.The overall rate of limb thrombosis waslow, with most limb occlusions occur-ring � 6 months of stent-graft insertion,and would probably have been evenlower had we been able to identify allhigh-risk cases for prophylactic adjunc-tive stenting. Limb occlusion denotes anunderlying problem with the graft,which if left untreated after thrombec-tomy or thrombolysis will lead to re-thrombosis. Postoperative imaging wasof little value in detecting impendinglimb occlusion. Based on these findings,we believe one should identify and stentany limbs that appear to be at risk forthrombosis, but this study lacks the datato predict which limbs need stenting.Authors’ Abstract

Sheehan MK, Ouriel K. Are type IIendoleaks after endovascular aneu-rysm repair endograft dependent? JVasc Surg 2006; 43(4):657–661.

• PURPOSE: Most clinical endpoints after endovascular aneurysm re-pair (EVAR) are endograft-specific, buttype II endoleaks have been assumed tobe an unavoidable consequence of therepair method and independent of thetype of endograft used. Some recent datahave suggested that the rate of type IIendoleaks may also be graft-dependent.We reviewed a large clinical experiencewith six endografts to determine the be-havior of type II endoleaks and whetherthey are graft-specific. METHODS: Allelective EVAR cases from five universityinstitutions from 1996 to 2003 were ret-rospectively analyzed. Endografts usedin �50 patients were excluded. En-doleaks were diagnosed and classifiedfrom contrast-enhanced computed to-mography (CT) scans by the treating sur-geons. Results of angiography and inter-ventions for endoleaks were tracked.The rate of type II endoleaks was com-pared among endografts at 1, 6, and 12months, and yearly thereafter. Statisticalsignificance was defined as P � .05. RE-SULTS: During the study period, 1909patients underwent elective EVAR andhad an adequate imaging follow-up atone of the specified time points. At 1month, the overall rate of type II en-doleak was 14.0% (range, 9.8% to 25.2%.)The Excluder had a significantly higherincidence of type II endoleaks at 1 monthbut was similar to most other grafts dur-ing longer follow-up. At 6 months, theoverall rate of type II endoleak was16.3% (range, 8.3% to 16.8%). The Talentand Lifepath had an apparent lower ini-

Abstracts • 1217Volume 17 Number 7

tial rate of type II leaks, but this was onlysignificant for the Talent at 6 monthscompared with Excluder, Zenith, andAncure, and at 1 year compared withExcluder and Zenith. No graft had along-term statistically significant differ-ence in the rate of type II endoleak for-mation. Intervention rates varied by in-stitution and graft type but in generalwere quite low. Of 25 successful inter-ventions (Ancure, 12; AneuRx, 8; Ex-cluder, 2; Lifepath, 2; Zenith, 1; Talent,0), 21 were performed during the firstyear. Interim spontaneous resolution,defined as a negative CT scan after a CTpositive for endoleak, was high, espe-cially in the first year. Resolution of typeII endoleaks occurred in 54 of 164 (33%)between 1 and 6 months, in 37 of 112(33%) between 6 and 12 months, in 20 of57 (35%) from 12 to 24 months, and in 5of 25 (20%) between 24 and 36 months.The various grafts had a nearly identicalpattern, but the rates were highest forthe Talent. Late appearing endoleaks,defined as a positive CT after a negativeCT, were frequent. At 6 months, 44(30%) of 147 type II endoleaks werenewly diagnosed. The rates were 37 of107 (35%) at 12 months, 15 of 56 (27%) at2 years and 5 of 20 (25%) at 3 years. Noconversions to open repair for type IIendoleaks were noted in the first 4 years.The thrombus burden could not be de-termined in this analysis. CONCLU-SIONS: Type II endoleaks occur innearly 15% of patients treated by EVAR.The early incidence varies only slightlywith graft type. The long-term preva-lence and clinical significance aremasked by different treatment patterns,spontaneous resolution, newly evidentendoleaks, and aneurysm size at initialtreatment.Authors’ Abstract

Greenberg RK, West K. Beyond theaortic bifurcation: branched endovas-cular grafts for thoracoabdominal andaortoiliac aneurysms. J Vasc Surg2006; 43(5):879–886; discussion 886–887.

• OBJECTIVES: To evaluate the useof novel technology to treat complex aor-tic aneurysms involving branches thatprovide critical end-organ blood supply.METHODS: A prospective study wasconducted in patients with thoracoab-dominal, suprarenal, or common iliacaneurysms (TAA, SRA, or CIA) at highrisk for open surgical repair. An endo-vascular graft using the Zenith platformwas customized to fit patient anatomy(TAA or SRA) and combined withJomed balloon-expandable stent-grafts.Prefabricated hypogastric branches wereused with a Zenith abdominal aortic an-eurysm (AAA) or Fluency self-expand-

ing fenestrated device in conjunctionwith a self-expanding stent-graft. Anal-yses were conducted in accordance withthe endovascular aneurysm reportingstandards document. Follow-up studiesoccurred at discharge, 1, 6, and 12months, and included computed tomog-raphy and duplex ultrasound scans, andflat plate radiography. RESULTS: Fiftypatients were treated (9 TAA, 20 SRA, 21CIA). The mean aneurysm size was 7.6cm (TAA), 7.2 cm (SRA), and 6.1 cmAAA size associated with a mean CIAsize of 3.8 cm. Bilateral CIA aneurysmswere present in 86% (18/21) of patientswith CIA aneurysms. Perioperative mor-tality was 2% (1/50) and resulted from amyocardial infarction after a plannedconduit and iliac endarterectomy re-quired for device access. Five late deathsoccurred (2 TAA, 2 SRA, 1 CIA), three ofwhich (2 TAA, 1 SRA) were aneurysmrelated. Failure to access internal iliacarteries occurred in three cases, and twolate hypogastric branch thromboses oc-curred. No visceral branches were lostacutely or occluded during follow-up.Sac shrinkage (�5 mm) was noted in65% of patients at 6 months and in allpatients (10/10) by 12 months. Therewere no ruptures or conversions, butnine patients required secondary inter-ventions. CONCLUSIONS: Branch ves-sel technology has made it technicallyfeasible to preserve critical end-organperfusion in the setting of CIA, SRA, andTAA aneurysms. The relatively lowacute mortality rate and lack of short-term branch vessel loss are encouragingand merit further investigation. Theseadvances have the potential to markedlydiminish the complications associatedwith conventional management of com-plex aneurysms.Authors’ Abstract

Fairman RM, Nolte L. Factors predic-tive of early or late aneurysm sac sizechange following endovascular repair.J Vasc Surg 2006; 43(4):649–656.

• OBJECTIVE: The purpose of thisstudy was to determine the relationshipbetween aneurysm sac size change at 1,6, 12, and 24 months and a set of 10independent “predictive” variables byusing a general linear model analysis.METHODS: In a multicenter trial, 351patients received the Zenith tri-modularbifurcated endograft. The predictivevariables used for this analysis were en-doleak by type, age, gender, smokingstatus, and the preprocedure variables ofmaximum aneurysm major diameter,minor neck diameter, proximal necklength, neck plaque/thrombus, and neckshape; and patent inferior mesenteric ar-tery at predischarge. The aneurysmchange was calculated as the difference

from the predischarge (� 7 days of im-plant) maximum aneurysm major diam-eter measurement to the maximum an-eurysm major diameter measurement atfollow-up examination periods of 1, 6,12, and 24 months. The same 10 predic-tive variables were used to assess theabsolute change in maximum aneurysmminor diameter and aneurysm area. Ad-ditionally, the percent change from pre-discharge was also assessed for the ma-jor diameter, minor diameter, andaneurysm area. RESULTS: None of theindependent variables were predictiveof absolute sac size change or percentchange at 1 month. At 6 months, thepresence of an endoleak (P � .01) andpreprocedure neck thrombus/plaque (P� .01) were significant predictors of ab-solute and relative aneurysm sizechange for all measurements (major di-ameter, minor diameter, and area) andwere more likely to be associated withless sac shrinkage or to have sac growth.Additionally, preoperative maximumaneurysm major diameter was a signifi-cant predictor for absolute change inarea (P � .01). Larger preprocedure an-eurysm diameters were more likely toexperience more shrinkage. The signifi-cant predictors of size change at 12months included preprocedure maxi-mum aneurysm major diameter, thepresence of endoleak at 12 months, pre-operative neck thrombus/plaque, andgender. At 24 months, significant predic-tors of aneurysm size change includedpreprocedure maximum aneurysm ma-jor diameter, endoleak at 24 months, andpreprocedure neck thrombus/plaque.When the longitudinal model was used,the presence of an endoleak, thrombus/plaque within the proximal neck at pre-procedure, and preprocedure maximumaneurysm major diameter were found tobe significantly related to the size of themaximum aneurysm major diameterover time. CONCLUSIONS: This studysupports the concept that early and latesac size change following EVAR is influ-enced by identifiable independent pre-dictive variables.Authors’ Abstract

Adam DJ, Fitridge RA. Late reinter-vention for aortic graft-related eventsand new aortoiliac disease after openabdominal aortic aneurysm repair inan Australian population. J Vasc Surg2006; 43(4):701–705; discussion 705–706.

• OBJECTIVE: To examine late rein-tervention rates for aortic graft-relatedevents and new aortoiliac disease afteropen abdominal aortic aneurysm (AAA)repair in an Australian population.METHODS: Interrogation of a prospec-tive computerized database identified1256 consecutive patients (1058 men, 198

1218 • Abstracts July 2006 JVIR

women; median age, 70 years; range, 40to 97 years) who survived open repair ofnonruptured (n � 957, group I) and rup-tured (n � 299, group II) infrarenal AAAin a single institution between January 1,1982 and December 31, 2003. Median(range) follow-up was 41 (1 to 261)months for group I and 30 (1 to 243)months for group II. RESULTS: In groupI, 33 patients (3.4%) underwent 38 latereinterventions: 20 patients (2.1%) foraortic graft-related events at a median(range) interval of 36 (1 to 94) monthsafter the index AAA repair, with a 30-day mortality rate of 15%; and 13 pa-tients (1.4%) for new aortoiliac disease ata median (range) interval of 33 (3 to 207)months, with 30-day mortality of 8%. Ingroup II, 15 patients (5%) underwent 16late reinterventions: 10 patients (3.3%)for aortic graft-related events at a me-dian (range) interval of 5 (2 to 112)months, with a 30-day mortality of 10%;and five patients (1.7%) for new aor-toiliac disease at a median (range) inter-val of 67 (39-105) months, with a 30-daymortality of 40%. There was no signifi-cant difference in the late reinterventionrate between the groups: group I, 33(3.4%) of 957 vs group II, 15 (5%) of 299(P � .23). For all patients, the estimatedsurvival at 1, 3, 5 and 10-years was 90%,79.4%, 66.4%, and 31.6%, respectively;estimated survival free from reinterven-tion at 1, 3, 5 and 10-years was 98.7%,97.1%, 95.1%, and 91.9%, respectively.CONCLUSIONS: These data demon-strate, for the first time, that open AAArepair has excellent long-term durabilityin an Australian population and the re-sults compare favorably with previousreports from North America and Eu-rope. These data represent an importantbenchmark for comparison of the resultsof endovascular AAA repair in this pa-tient population.Authors’ Abstract

Hobo R and Buth J. Secondary inter-ventions following endovascular ab-dominal aortic aneurysm repair usingcurrent endografts. A EUROSTAR re-port. J Vasc Surg 2006; 43(5):896–902.

• OBJECTIVE: The purpose of thisstudy was to evaluate the need for sec-ondary interventions after endovascularabdominal aortic aneurysm repair withcurrent stent-grafts. METHODS: Studiedwere data from 2846 patients treatedfrom December 1999 until December2004. The data were recorded from theEUROSTAR registry. The only patientsstudied were those with a follow-up ofat least 12 months or until they had asecondary intervention within the first

12 months. The cumulative incidences ofsecondary transabdominal, extra-ana-tomic, and transfemoral interventionsduring follow-up (after the first postop-erative month) were investigated. RE-SULTS: A secondary intervention wasperformed in 247 patients (8.7%) at amean of 12 months after the initial pro-cedure within a follow-up period of amean of 23 � 12 months. Of these, 57(23%) transabdominal, 43 (16%) in-volved an extra-anatomic bypass, and147 (60%) were by transfemoral ap-proach. The cumulative incidence of sec-ondary interventions was 6.0%, 8.7%,12%, and 14% at 1, 2, 3, and 4 years,respectively. This corresponded with anannual rate of secondary interventionsof 4.6%, which was remarkably lowerthan in a previously published EURO-STAR study of patients treated before1999. Type I endoleaks (33% of proce-dures), migration (16%), and rupture(8.8%) were the most frequent reasonsfor secondary transabdominal interven-tions. Graft limb thrombosis was the in-dication for extra-anatomic bypass(60%). Type I endoleak (17%), type IIendoleak (23%), device limb stenosis(14%), thrombosis (23%), and device mi-gration (14%) were the most frequentreasons for secondary transfemoral in-terventions. Operative mortality washigher after secondary transabdominalinterventions (12.3%, P � .007) com-pared with transfemoral interventions(2.7%). Overall survival was lower in pa-tients with secondary transabdominal (P� .016) and extra-anatomic interventions(P � .0001) compared with patientswithout a secondary intervention. CON-CLUSION: Although the incidence ofsecondary interventions after endovas-cular aneurysm repair has substantiallydecreased in recent years, continuingneed for surveillance for device-relatedcomplications remains necessary.Authors’ Abstract

Fillinger M. Three-dimensional anal-ysis of enlarging aneurysms after en-dovascular abdominal aortic aneurysmrepair in the Gore Excluder Pivotalclinical trial. J Vasc Surg 2006; 43(5):888–895.

• OBJECTIVE: Recent reports haveraised concern about the percentage ofenlarging abdominal aortic aneurysms(AAAs) after endovascular repair withthe Gore Excluder device. As part of theinvestigation into this issue, a morpho-logic analysis was performed on enlarg-ing aneurysms in the Excluder Pivotalclinical trial. METHODS: Computed to-mographic scans were evaluated on all

patients identified with enlarging aneu-rysms (5-mm increase by Core labora-tory or site) and at least 4 years of fol-low-up in the Excluder Pivotal clinicaltrial. Three-dimensional reconstruction,a set of 24 standard morphologic mea-surements, and analysis of potential en-largement mechanisms were performed.RESULTS: Of 112 trial patients with 4years of follow-up, 38 AAAs (34%) wereidentified as enlarging. Data were ob-tained from 196 computed tomographicscans (the mean interval was 47 monthsfrom first to last scan). Of the 158 scanswith a prior scan for comparison, 41%demonstrated growth relative to the ini-tial scan by diameter criteria, but 79%demonstrated growth relative to the ini-tial scan by 3-dimensional volume crite-ria (P � .0001 vs diameter; �2 analysis).This difference was most evident at earlytime points: at 1 year, diameter criteriaindicated that 8% of these AAAs wereenlarging, but 56% were already enlarg-ing by volume criteria. On average, en-largement was detected by volume 18months before it was detected by diam-eter (P � .0001), and at a smaller diam-eter (55 � 1 mm vs 60 � 1 mm; P �.0001). Only 19% of scans (39% of pa-tients) had apparent endoleaks. Scanswith apparent endoleaks demonstrateda greater interval rate of growth as com-pared with those without apparent en-doleak (3.6 � 0.8 mm vs 1.9 � 0.3 mm [P� .02] by diameter; 23 � 4 cm3 vs 11 � 1cm3 [P � .001] by volume). Although theetiology of enlargement may be endo-tension or device permeability in up to74% of patients, other potential causes ofaneurysm enlargement included neckapposition length less than 15 mm (15patients; 39%), large aortic diameter rel-ative to device (18%), large iliac diameter(5%), and iliac apposition length lessthan 15 mm (20%). Multiple potentialetiologies of enlargement were presentin 53% of AAAs. CONCLUSIONS: Theetiology of aneurysm enlargement in theExcluder Pivotal trial is likely multifac-torial, including endoleak, inadequateattachment site length, and endotensionor device permeability. Even by conser-vative criteria, a substantial percentageof aneurysm growth with the originaldevice is likely due to material perme-ability. Three-dimensional volume crite-ria detected aneurysm enlargementmore frequently, at a smaller diameter,and on average 18 months sooner thanstandard diameter criteria, thus suggest-ing a role in further investigation of thisissue.Authors’ Abstract

Abstracts • 1219Volume 17 Number 7