Comparison of Color Duplex Ultrasound andComputed Tomography Scan forSurveillance after Aortic Endografting

Sandrine Pages, MD,1 Jean-Pierre Favre, MD,2 Alexis Cerisier, MD,3

Seeven Pyneeandee, MD,1 Christian Boissier, MD,4 and Charles Veyret, MD,1

Saint Etienne, France

Endovascular repair of abdominal aortic aneurysms (AAA) requires regular surveillance for earlydetection of endograft failure. CT scanning is the gold standard surveillance procedure. Thepurpose of this study was to assess the reliability of color duplex ultrasound (CDU) in compari-son to CT scanning for detection of endoleaks and changes in aneurysmal diameter. FromNovember 1996 to September 1999, a total of 41 patients treated by aortic endografting under-went regular surveillance with both CT scanning and CDU. There were 39 men and 2 womenwith a mean age of 71 years (range, 50-83). Endovascular treatment involved deployment of astraight aorto-aortic stent in 6 cases, bifurcated stent in 33, and aorta-to-unilateral iliac arterystent in 2. Stent deployment failed in one case; the procedure was conversion to open surgery.Primary or secondary endoleaks were detected in 17 patients (42%). Our findings indicated thatCDU is less reliable than the CT scan for detection of endoleaks, but that reliability of CDU forsurveillance of aneurysmal diameter is fair.

INTRODUCTION

Since the initial report by Parodi et al.1 in 1991 onendovascular repair of abdominal aortic aneurysm(AAA), the popularity of this technique has grown.Good short-term results have been reported by sev-

eral teams using diverse endografts.2-7 However,mid- and long-term outcome is still unclear andnumerous problems must be solved before endo-vascular treatment can supplant open surgical re-pair.

Graft failure is the most common short and mid-term complication after aortic endografting. Themain cause of graft failure is leakage between theprosthesis and arterial wall. The incidence of en-doleak has ranged widely from 4 to 40%, depend-ing on the stent device used, patient selection, andduration of stent placement.2,3,8-13 While primaryendoleaks are most frequent, long-term develop-ment of this complication is also possible not onlybecause of stent deterioration but also because ofchanges in aneurysmal morphology after treat-ment. It has now been shown that aneurysmal di-ameter decreases after exclusion and that rupture isa potential complication in patients with persistentendoleaks.14,15 On the basis of these observations,

1Radiology Department, Saint Etienne University Hospital Center,Saint Etienne, France.

2Cardiovascular Surgery Department, Saint Etienne University Hos-pital Center, Saint Etienne, France.

3Cardiology Department, Saint Etienne University Hospital Center,Saint Etienne, France.

4Vascular Laboratory Department, Saint Etienne University Hospi-tal Center, Saint Etienne, France.

Correspondence to: J.-P. Favre, MD, Service de Chirurgie Cardio-vasculaire, Hopital Nord CHU Hopitaux de Saint-Etienne, 42055 Saint-Etienne Cedex 2, France. E-mail: favrejp@aol.com.

Ann Vasc Surg 2001; 15: 155-162DOI: 10.1007/s100160010065© Annals of Vascular Surgery Inc.Published online: March 1, 2001

155

screening and treatment of endoleaks have becomeessential issues in endovascular AAA treatment,and life-long surveillance is needed.

The criteria for successful endovascular treat-ment are decrease in maximum aneurysmal diam-eter, absence of endoleak, stability of neck diam-eters, and absence of stent-related problems. Threeof these criteria can be analyzed by spiral CT scan-ning, which is currently the gold standard tech-nique for surveillance of aneurysms after endovas-cular treatment. However, this technique has sev-eral drawbacks, including use of contrast dye andradiation. These problems could be avoided by ul-trasonongraphic surveillance. The purpose of thisstudy was to compare color duplex ultrasonography(CDU) and CT scanning for follow-up of patientstreated by endovascular aortic endografting.

PATIENTS AND METHODS

From November 1996 to September 1999, a total of190 patients presenting with infrarenal AAA werereferred for elective treatment at our institute.Work-up included assessment aneurysmal mor-phology by CT scan and arteriography to select pa-tients for endovascular treatment. In 41 of thesepatients (21.6%), anatomical findings were com-patible with stent-graft placement. There were 39men and 2 women with a mean age of 71 years(range, 50 to 83). American Society of Anesthesi-ologists (ASA) classification is shown in Table I.Thirteen patients had a history of previous laparot-omy.

Mean preoperative aneurysmal diameter deter-mined by CT scan was 55 ± 9 mm (range; 40 to 90mm). The proximal neck of the aneurysm was lo-cated below the renal arteries in all cases. Maxi-mum proximal neck diameter was 28 mm andminimal length was 15 mm. In seven cases the dis-tal neck was located at the end of the aorta. Maxi-mum distal neck diameter in these cases was 26 mmand minimal length was 10 mm. These cases weretreated by placement of a straight aorto-aortic stentin six cases and an aorto-to-unilateral iliac stent

associated with crossover femorofemoral bypass inone case. In 34 cases the distal neck was not locatedin the aorta. These cases were treated by placementof a bifurcated stent in 33 cases and an aorto-to-unilateral iliac stent associated with crossoverfemorofemoral bypass in 1 case. The four differentstent models used are listed in Table II.

All procedures were performed in the operatingroom under general anesthesia using a digital im-aging system (BV29, Philips, Eindhoven, The Neth-erlands). Depending on the type of stent device, asingle or double femoral or iliac approach was used.In one case, endograft deployment failed because ofextensive calcification of the iliac arteries. Opensurgery was undertaken immediately. In the re-maining 40 cases, stent placement was achieved foran immediate success rate of 97.5%. There were nopostoperative deaths. The mean duration of hospi-talization was 4 days. The only postoperative com-plication was one infection in Scarpa triangle thatresponded to local treatment.

Postoperative surveillance included plain ab-dominal roentgenography, CT scan, and CDU.These procedures were performed prior to dis-charge and at 3, 6, 12, 24, and 30 months. Studyfocused on detection of endoleaks, measurement ofmaximum AAA diameter, and assessment of metal-lic stent structure. CT scan and CDU examinationwere performed by different operators at differentlocations. The second operator had no knowledgeof the results of the first examination. All examina-tions, i.e., CT scans and videotaped CDU proce-dures, were reviewed by an independent radiolo-gist.

Three consecutive spiral CT scan acquisitionswere performed using a Somatom Plus S® system(Siemens, Erlongen, Germany). The first scan wasdone without contrast dye with a section thicknessof 10 mm to locate the superior mesenteric artery(SMA) and hypogastric arteries. Next the aorta wasvisualized from the SMA to the hypogastric arteriesusing two adjoining spiral CT scans after injection ofcontrast dye. Collimation settings for aortic visual-

Table I. ASA classification of 41patients treated by aorticendografting

Classification Patients

ASA 2 27ASA 3 12ASA 4 2

ASA, American Society of Anesthesiologists.

Table II. Stent devices used for aorticendografting

Straight BifurcatedAorto-uni-iliac Total

EVT 6 5 1 12Vanguard 0 25 0 25AneuRx 0 1 1 2Endologix 0 2 0 2Total 6 33 2 41

156 Pages et al. Annals of Vascular Surgery

ization were between 3 and 5 mm and pitch was 1.CDU was performed by three qualified angiologistsusing an Apogee® 800PLUS ultrasound system(ATL, Philips, Eindhoven, The Netherlands) with a3.5 MHz curved array transducer. The abdominalaorta was visualized from the celiac trunk to thehypogastric arteries first in the transverse plane andthen in the longitudinal plane in the B-mode andcolor Doppler mode. Continuous spectral analysiswas performed if color Doppler findings suggestedthe presence of an endoleak.

Endoleaks were defined as persistent blood flowor uptake of contrast between the stent graft andwalls of the aneurysmal sac. On CT scans, endoleakswere characterized by extravasation of contrast dyebetween the prosthesis and aneurysmal wall. OnCDU the characteristic feature was detection of acolor and spectral signal outside the limits of theprosthesis. Endoleaks detected during the first 30days after endografting were considered primary.Beyond that time, endoleaks were classified as sec-ondary. Type I endoleaks were defined as leaks re-sulting from leakage around the proximal or distalneck of the stent, through the stent wall, or at junc-tions between modular stents. Type II endoleakswere defined as leaks resulting from recirculation inthe aneurysmal sac supplied by collateral vesselsfrom lumbar arteries and/or inferior mesenteric ar-tery.12

Primary endoleaks were observed in seven pa-tients (17%). The underlying mechanism wasproximal neck leakage in two cases, distal neckleakage in two, transprosthetic leakage in two, andcollateral recirculation in one case. Four of theseendoleaks resolved spontaneously, two were suc-cessfully treated by endovascular techniques—embolization of the hypogastric artery and iliac ar-tery stent—and one persisted. Secondary endoleakswere observed in 10 patients (25%). The delay fordevelopment of secondary endoleaks ranged from 1to 30 months after stent placement. Four of theseendoleaks resolved spontaneously, one was unsuc-cessfully treated by the endovascular route, andfour persisted. One patient died suddenly of an un-known cause 6 months after placement, presum-ably with persistent endoleak. Of the 40 patients inwhom aortic endografting was successful, 17 (42%)developed an endoleak within a mean delay of 12months after stent placement. Six of these leakswere persistent at the end of study (Table III).

On CT scan, the maximum diameter of the an-eurysm was defined as the largest diameter mea-sured regardless of the position of the aortic axis,including the thickness of the wall. Fixed anatomi-cal landmarks, e.g., vertebrae, calcification, and kid-neys, were used as landmarks to allow reproduciblemeasurements from one examination to thenext.16,17 On CDU, the largest anterioposterior or

Table III. Description and outcome of endoleaks

Case Type

Delay forappearance(months) Treatment

Resolution(months)

Follow-up(months)

Primary endoleaks1 II 0 0 Yes (12) 122 I 0 0 No 63 I 0 0 Yes (12) 124 I 0 1a Yes (8) 185 I 0 1a Yes (12) 246 I 0 0 Yes (1) 37 I 0 0 Yes (1) 18

Secondary endoleaks8 I 3 0 No 39 II 3 0 No 6

10 I 3 0 Yes (24) 2411 I 6 0 Yes (12) 1212 II 6 1a No 3013 II 3 0 No 314 II 12 0 No 1215 II 3 0 Yes (12) 1216 I 1 0 No 617 I 6 0 Yes (24) 24

aTreatment of endoleaks: placement of another endograft distally (n = 4), embolization of acollateral hypogastric artery (n = 5); embolization and placement of another endograft distally(n = 12).

Vol. 15, No. 2, 2001 CDU and CT surveillance after aortic endografting 157

transverse diameter was used.18,19 The patientpopulation was divided into two groups; group I (n= 23) included patients with no endoleak and groupII (n = 17) included patients with endoleak.Changes in maximum AAA diameter between twofollow-up examinations were analyzed on bothCDU and CT scan. Diameter changes observed byCDU were compared with diameter changes ob-served by CT scanning, considered the gold stan-dard.

Statistical analysis was performed using Stat-view®. Differences in diameters measured by CTscan and CDU were compared using the chi-squared test for discrete variables and the Studentt-test for continuous variables. A p value of <0.05was considered significant. Sensitivity, specificity,positive predictive value (PPV), and negative pre-dictive value (NPV) were calculated for detection ofendoleaks and diameter variations. The Kappa con-cordance test was used to compare results obtainedby CDU and CT scan.

RESULTS

All patients underwent regular follow-up but bothprocedures were not always performed at eachscheduled control examination. In a total of 137control examinations scheduled, 123 CT scans and121 CDU were performed. Both CDU and CT scanwere performed to detect endoleak and changes inmaximum AAA diameter in a total of 109 controlexamination. Of these 109 procedures, 97 were re-corded for later review. B-mode imaging was satis-factory for interpretation in 91 of these 97 proce-dures. In six cases, B-mode images were uninter-pretable because of the presence of intestinal gas. In55 cases, spectral study was necessary to confirm ordeny suspicion of an endoleak based on color Dopp-ler findings (Table IV). All CT scans were completedin accordance with the study protocol and were re-viewed.

Detection of Endoleaks

Endoleak was diagnosed in 17 patients (42%) usingone or more examinations during follow-up. Of the109 control procedures , the endoleak was detectedby CT scan in 29 cases (27%). In 14 of these 29cases, the endoleak was also visualized by CDU. Inaddition, CDU visualized five endoleaks that werenot detected by CT scan (Table V). Of the five false-positive endoleaks visualized by CDU, one observedat two successive examinations corresponded to aabnormal systolic motion movement from the pos-terior side of the proximal segment of an EVT®endostent. It was difficult to be sure whether thisfinding corresponded to an endoleak. Two othercases corresponded to type II endoleaks. The re-maining case was an endoleak of undeterminedcause diagnosed at 3 months and not present on thefollowing control. Of the 15 false-negative en-doleaks not visualized by CDU, 1 patient had intes-tinal gas that prevented proper visualization, 8 hadtype II endoleaks located near the distal end of theaorta, and 6 had type I endoleaks involving the dis-tal neck in 5 cases and the proximal neck in 1 case.

With reference to CT scan as the gold standard,the sensitivity and specificity of CDU were 48% and94%, respectively. The PPV was 74% and the NPVwas 81%. Concordance between the two proce-dures seemed fair with a Kappa coefficient of 0.47(0.41 to 0.60).

Detection of Diameter Changes

As shown in Table VI, the mean maximum AAAdiameter for the 109 examinations involving bothprocedures was 52 ± 10 mm (range; 30 to 96 mm)for CT scanning and 46.5 ± 9 mm (range; 26 to 81mm) for CDU. This difference was statistically sig-nificant (p < 0.0001) for the overall population andfor the different subgroups regardless of the delayafter placement. In group I, which included patientswithout endoleaks, mean maximum AAA diametergradually decreased during the study. Conversely,

Table IV. Color duplex ultrasound proceduresperformed during follow-up of 40 patients treatedfor AAA by endografting

Scheduled 137Actually performed 121Concurrent with CT scan 109Tape recorded 97Interpreting B-mode 91Interpreting color ultrasound 91Spectral analysis 55

Table V. Comparison of CDU and CT scan fordiagnosis of endoleaks after endograft treatmentof 40 AAA

CDU

CT scan Endoleak No endoleak Total

Endoleak 14 15 29No endoleak 5 75 90Total 19 80 109

158 Pages et al. Annals of Vascular Surgery

in group II, the patients with endoleaks, meanmaximum AAA diameter gradually increased(Table VII). These finding were the same whateverthe technique used, CT scan or CDU.

Intercontrol variation in AAA diameter was ana-lyzed for comparison of findings at 0 and 3 months,3 and 6 months, 6 and 12 months, and 12 and 24months. Analysis of diameter changes was per-formed with respect to the 64 intervals for whichcomparison between CT scan and CDU was pos-sible. Diameter decreased 38 times based on CT scanfindings and 32 times based on CDU. Increase indiameter was noted more frequently with CDUthan with the CT scan—i.e., 10 vs. 9 times (TableVIII). Using the CT scan as the reference technique,the sensitivity and specificity of CDU for demon-strating no change in AAA diameter were 88% and76%. The PPV was 72% and the NPV was 75% withan accuracy of 81% (Table IX). Concordance be-tween the two procedures seemed fair with a Kappacoefficient of 0.62 (0.61 to 0.80).

DISCUSSION

Endografting for AAA does not achieve exclusion ofthe aneurysm in all cases. Failures result from per-sistence of high pressure between the stent and an-eurysm, usually due to an endoleak. Although mostseries20,21 have reported endoleaks, incidence hasbeen highly variable. In the initial study using EVTendostents, Moore and Rutherford8 reported en-doleaks in 44% of cases; White and May,9 who usedseveral different stent models, reported endoleaksin 10% of cases; and Nothenius et al.22 reportedendoleaks in 4% of cases at 1 year after deploymentof the AneuRx® stent. In our series, we detectedendoleaks in 42% of patients during the first year offollow-up.

Endoleaks are usually associated with nonregres-sion or even expansion of the AAA.23,24,25,26. Insome cases, nonregression of the aneurysm is ob-

served without the presence of an endoleak. Thesefindings mandate close surveillance. The endoten-sion concept proposed by Gilling-Smith et al. 27

could explain nonregression without an endoleak.With or without an endoleak, an aneurysm thatdoes not decrease in size during follow-up cannotbe considered to be adequately treated. Under theseconditions, regular monitoring is required until de-finitive exclusion of the AAA is achieved.

The reference technique for surveillance of en-dografts is CT scan. 8,20,28,29 It allows detection ofendoleaks, measurement of diameter of the AAAand its necks, and, to some extent, assessment ofmechanisms underlying endoleakage.23,25 How-ever, even CT scanning may not be completely ad-equate. According to Jaakkola et al.,30 diametermeasurements by CT scan are subject to method-ological error and interobserver variability. Mea-surements using a cursor on a small image can beoff by several millimeters. According to Lederle etal.,16 interobserver variations are <2 mm in 93% ofcases but can be as high as 5 mm.

The usefulness of CDU in comparison with CTscan is unclear since there are few studies that de-fine its indications.31 Potential benefits includegood reproducibility and avoidance of risks andcomplications associated with iodinated contrastdyes and ionizing radiation. Lanne et al.18 wrotethat ultrasound was accurate enough to detect di-ameter changes of 2 mm or less. Jaakkola et al.30

stated that the reliability of the CT scan and CDUwas equivalent and cautioned that intra- and in-terobserver variations had to be taken into accountusing both modalities. Lederle et al.16 reported thatultrasonographic measurements were less accuratethan CT scan measurement. The disadvantages ofCDU are well known, since interpretability dependson the echogenicity of the patient, system quality,and operator skill.

In a recent comparative study, Cao et al.32 vali-dated the utility of CDU for detection and assess-ment of endoleaks. In a series of 198 examinationsusing both CDU and CT scan as the reference tech-nique, ultrasound detected some false negatives(sensitivity, 91.7%; NPP, 99.4%) but overestimatedthe incidence of endoleaks. In a study comparingCDU and CT scan for detection of endoleaks afterdeployment of EVT endostents, Sato et al.19 dem-onstrated excellent sensitivity (97%) but only fairspecificity (74%). The same authors also empha-sized several technical problems that greatly im-paired the reproducibility of CDU. They concludedthat CDU was reliable in only 19% of cases. In ourexperience, the sensitivity and specificity of CDUwere 48% and 93%, respectively. The most likely

Table VI. Mean maximum diameter (in mm)measured by CT scan and CDU duringsurveillance after endograft treatment of 40 AAA

Total Group Ia Group IIb

CT scan 52 ± 10 48 ± 6 61.5 ± 14CDU 46.5 ± 9 43.5 ± 8 53.5 ± 10p <0.0001 <0.0001 <0.0001

Values are expressed in millimeters with standard deviation.aGroup I: no endoleak.bGroup II: endoleak.

Vol. 15, No. 2, 2001 CDU and CT surveillance after aortic endografting 159

explanation for this difference in results is that weconsidered findings analyzable even if they did notpresent all the criteria proposed by Sato et al.19

In our series, CDU detected five endoleaks thatwere not observed on CT scan. One explanation forthese false positives is that all of our contrast-enhanced CT scans were performed immediately af-ter injection without delayed sequences. Currentevidence suggests that routine use of delayed se-quences can improve the reliability of the CTscan.33 Nonetheless, the finding that CDU can de-tect endoleaks missed by a CT scan suggests that

combined use of these two modalities could im-prove the quality of surveillance after aortic endo-grafting. Conversely, in >50% of cases in our series,CDU failed to detect endoleaks detected by CT scan.The explanation for this high false-negative rate isunclear. This finding raises questions not onlyabout the value of CDU for endostent surveillancebut also about the conditions under which the pro-cedure was performed. In this regard, we were sur-prised to observe that most undetected endoleakswere located near the aortic bifurcation, which isusually readily visualized using CDU.

As pointed out by Sato et al.,19 CDU could enablea better understanding of the mechanisms under-lying some endoleaks. Specifically, CDU may pro-vide more precise analysis in cases involving en-doleaks due to collateral recirculation. However, ac-cording to Cao et al.,32 the ability of CDU topinpoint the source of leakage is low, with a sensi-tivity of only 60.7%. Other authors31,34 have re-ported that use of contrast dyes can improve theperformance of CDU in this regard. One recentstudy by a group in Liverpool compared CT scan-ning and CDU with or without Levovist for detec-tion of endoleaks after deployment of an endograft.A total of 20 comparative examinations were car-ried out in 18 patients. CT scan, which was used asthe reference modality as in previous comparativestudies, detected three endoleaks. CDU without Le-vovist detected only one of these leaks whereasCDU with Levovist detected all three. In addition,CDU with Levovist revealed six endoleaks not de-tected by CT scan, including one case involving anunexplained increase in aneurysm diameter. De-spite the small size of their study population, theauthors concluded that contrast-enhance ultra-sound could become the gold standard modality fordetection of endoleaks.

Table VII. Comparison of mean maximum diameter measured preoperatively and postoperatively byCT scan and CDU

Group Ia Group IIb

CT scan CDU Difference CT scan CDU Difference

Preoperative 51.5 ± 8 47.5 ± 6 4 55.5 ± 10 51 ± 10 4.56 months 48 ± 6 41 ± 6 7 53 ± 11 47 ± 7 612 months 44 ± 6 38 ± 5 6 52.5 ± 10 48.5 ± 7 4Differencec 7.5 9.5 3 2.5

Values are expressed in millimeters with standard deviationaGroup I: no endoleak.bGroup II: endoleak.cDifference between preoperative and 12-month values (mean value).

Table VIII. Diameter changes noted by CT scanand CDU during follow-up after endografttreatment of 40 AAA

CT scan CDU

Increase in diameter 9 10No change in diameter 17 22Decrease in diameter 38 32Total 64 64

Values are expressed in number of aneurysms examined.

Table IX. Changes in aneurysm diameter at twosuccessive control examinations using both CDUand CT scan

CDU

CT scan

Increaseor nochange

Decreasein diameter Total

Increase or no change 23 3 26Decrease in diameter 9 29 38Total 32 32 64

Values are expressed in number of aneurysms examined.

160 Pages et al. Annals of Vascular Surgery

Based on current knowledge, the most reliableproof of successful AAA exclusion is follow-up evi-dence of aneurysm size reduction.35 A decrease inaneurysm diameter is sometimes accompanied bylongitudinal aneurysm shrinkage that can lead todisplacement of the stent with kinking of itsbranches or disconnection of modular units, result-ing in secondary endoleaks.2,36 Nonregression of ananeurysm after stent deployment must be consid-ered an indication that exclusion has not beenachieved. CT scanning appears to be a reliable tech-nique for monitoring diameter since it allows mea-surements to be made on the same plane in succes-sive procedures. However, changes in aneurysmmorphology during follow-up are irregular, withdecreases being observed in some zones and nochange in others.23,37 Thus surveillance of the di-ameter of a single area is not necessarily sufficient.A more reliable approach could involve calculationof aneurysmal volume based on data from severalsections.7

Comparison of diameter measurements obtainedby CT scan and CDU is difficult because of differ-ences in the measurement techniques.16,30 A majordifference is that, unlike CDU, measurements usinga CT scan include the full thickness of the wall. Thisdiscrepancy could account for the differences com-monly reported between preoperative and postop-erative diameters.16 If a CT scan was used to evalu-ate aneurysm diameter initially, it is difficult toswitch to CDU during follow-up. However, as pre-viously noted by Malina et al.,3 there is a linearrelationship between diameter measurements in CTscanning and CDU. Thus the reliability of CDU issufficient for detection of an increase in diameter,provided that a preoperative ultrasound procedureis available for reference. In our experience, diam-eter measurements did not appear to be correlatedwith the technique used regardless of whether anendoleak was present. Our results also show goodconcordance between successive control examina-tions using CDU and CT scan for surveillance ofdiameter changes.

Based on our experience, we concluded thatCDU is less reliable than the CT scan for surveil-lance of AAA after endografting since the ability ofCDU to detect endoleaks is insufficient. However,CDU could reveal endoleaks missed by a CT scan,especially in association with the use of contrastdyes. Ultrasound data can also be helpful in under-standing the underlying mechanism of endoleakand seems reliable for surveillance of maximum an-eurysm diameter.

REFERENCES

1. Parodi JC, Palmaz JC, Barone HD. Transfemoral intralumi-nal graft implantation for abdominal aortic aneurysms. AnnVasc Surg 1991;5:491-499.

2. Matsumura JS, Moore WS. Clinical consequences of peri-prosthetic leak after endovascular repair of abdominal aorticaneurysm. J Vasc Surg 1998;27:606-613.

3. Malina M, Ivancev K, Chuter TAM, et al. Changing aneu-rysmal morphology after endovascular grafting: relation toleakage or persistent perfusion. J Endovasc Surg 1997;4:23-30.

4. Matsumura JS, Pearce WH, McCarthy WJ, et al. Reductionin aortic aneurysm size: early results after endovascular graftreplacement. J Vasc Surg 1997;25:113-123.

5. Walker ST, Macierewicz J, Elmarasy NM, et al. A prospectivestudy to assess changes in proximal aortic neck dimensionsafter endovascular repair of abdominal aortic aneurysms. JVasc Surg 1999;29:625-630.

6. Broeders AMJ, Blankensteijn JD, Gvakharia A, et al. Theefficacy of transfemoral endovascular aneurysm manage-ment: a study on size changes of the abdominal aorta duringmid-term follow-up. Eur J Endovasc Surg 1997;14:84-90.

7. Balm R, Kaatee R, Blankensteijn JD, et al. CT-angiographyof abdominal aortic aneurysms after transfemoral endovas-cular aneurysm management. Eur J Vasc Surg 1996;12:182-188.

8. Moore WS, Rutherford R. Transfemoral endovascular repairof abdominal aortic aneurysm: results of the North Ameri-can EVT phase 1 trial. J Vasc Surg 1996;23:543-553.

9. May J, White GH, Yu W, et al. Concurrent comparison ofendoluminal versus open repair in the treatment of abdomi-nal aortic aneurysms: analysis of 303 patients by life tablemethod. J Vasc Surg 1998;27:213-221.

10. Blum U, Voshage G, Lammer J, et al. Endoluminal stentgrafts for infra-renal abdominal aneurysms. N Engl J Med1997;336:13-20.

11. Raithel D. Surveillance des patients apres traitement endo-vasculaire ou conventionnel des anevrysmes de 1’aorte ab-dominale. J Mal Vasc 1998;23:390-392.

12. White GH, Yu W, May J, et al. Endoleak as complication ofendoluminal grafting of abdominal aortic aneurysms: classi-fication, incidence, diagnosis, and management. J EndovascSurg 1997;4:152-168.

13. Miahle C, Amicabile C, Becquemin JP. Endovascular treat-ment of infrarenal abdominal aneurysms by the Stentor sys-tem: preliminary results of 79 cases. J Vasc Surg1997;26:199-209.

14. Torsello GB, Klenk E, Kasprzak B, et al. Rupture of abdomi-nal aortic aneurysm previously treated by endovascularstentgraft. J Vasc Surg 1998;28:184-187.

15. Alimi YS, Chafke N, Rivoal E, et al. Rupture of an abdominalaortic aneurysm after endovascular graft placement and an-eurysm size reduction. J Vasc Surg 1998;28:178-181.

16. Lederle FA, Wilson SE, Johnson GR, et al. Variability inmeasurement of abdominal aortic aneurysms. J Vasc Surg1995;21:945-952.

17. May J, White GH, Yu W, et al. A prospective study ofchanges in morphology and dimensions of abdominal aorticaneurysms following endoluminal repair: a preliminary re-port. J Endovasc Surg 1995;2:343-347.

18. Lanne Y, Sandgren T, Mangell P, et al. Improved reliabilityof ultrasonic surveillance of abdominal aortic aneurysms.Eur J Endovasc Surg 1997;13:149-153.

19. Sato DT, Goff CD, Gregory RT, et al. Endoleak after aortic

Vol. 15, No. 2, 2001 CDU and CT surveillance after aortic endografting 161

stent graft repair: diagnosis by color duplex ultrasound scanversus computed tomography scan. J Vasc Surg1998;28:657-663.

20. Becquemin JP, Lapie V, Favre JP, et al. Mid-term results ofa second generation bifurcated endovascular graft for ab-dominal aortic aneurysm repair: the French Vanguard trial.J Vasc Surg 1999;30:209-218.

21. Zarins CK, White RA, Schwarten D, et al. AneuRx stent graftversus open surgical repair of abdominal aortic aneurysms:multicenter prospective clinical trial. J Vasc Surg1999;29:292-308.

22. Nolthenius RP, Berg JC, Biasi GM, et al. Endoluminal repairof infrarenal abdominal aortic aneurysms using a modularstent-graft: one-year clinical results from a European mul-ticenter trial. Cardiovasc Surg 1999;7:503-507.

23. Sonesson B, Malina M, Ivancev K, et al. Dilatation of infra-renal aneurysm neck after endovascular exclusion of ab-dominal aortic aneurysm. J Endovasc Surg 1998;5:195-200.

24. Resch T, Ivancev K, Lindh M, et al. Persistent collateral per-fusion of abdominal aortic aneurysm after endovascular re-pair does not lead to progressive change in aneurysm diam-eter. J Vasc Surg 1998;28:242-249.

25. Matsumura JS, Chaikof EL. Continued expansion of aorticnecks after endovascular repair of abdominal aortic aneu-rysms. J Vasc Surg 1998;28:422-431.

26. Umscheid T, Stelter WJ. Time-related alterations in shape,position, and structure of self- expanding, modular aorticstent-grafts: a 4-year single-center follow-up. J EndovascSurg 1999;6:17-32.

27. Gilling-Smith G, Brennan J, Harris P, et al. Endotension af-ter endovascular aneuvrysm repair: definition, classificationand strategy for surveillance and intervention. J EndovascSurg 1999;6:305-307.

28. Jacobowitz GR, Rosen RJ, Riles TS. The significance and

management of the leaking endograft. Semin Vasc Surg1999;12:199-206.

29. Tillich M, Hausegger KA, Tiesenhausen K, et al. Helical CTangiography of stent-grafts in abdominal aortic aneurysms:morphologic changes and complications. Radiographics1999;19:1573-1583.

30. Jaakkola P, Hippelainen M, Farin P, et al. Interobserver vari-ability in measuring the dimensions of the abdominal aorta:Comparision of ultrasound and computed tomography. EurJ Endovasc Surg 1996;12:230-237.

31. McWilliams RG, Martin J, White D, et al. Use of contrast-enhanced ultrasound in follow-up after endovascular aorticaneurysm repair. J Vasc Interv Radio 1999;10:1107-1114.

32. Cao, P, Zannetti, S, Verzini, F, et al. Imagerie postoperatoireapres endoprothese pour anevrysme de l’aorte. In:Branchereau, A, Jacobs, M, eds. Traitement chirurgical etendovasculaire des anevrysmes aortiques. Armonk, NY: Fu-tura Publishing, 2000, pp 119-126.

33. Schurink GWH, Aarts NJM, Wilde J, et al. Endoleakage afterstent-graft treatment of abdominal aneurysm. Implicationson pressure and imaging: an in vitro study. J Vasc Surg1998;28:234-241.

34. Heilberger P, Schunn C, Ritter W, Weber S, Raithel D. Post-operative color flow duplex scanning in aortic endografting.J Endovasc Surg 1997;4:262-271.

35. Beebe HG. Imaging modalities for aortic endografting. J En-dovasc Surg 1997;4:111-123.

36. Harris P, Brennan J, Martin J, et al. Longitudinal aneurysmshrinkage following endovascular aortic aneurysm repair: asource of intermediate and late complication. J EndovascSurg 1999;6:11-16.

37. Matsumura JS, Chaikof EL. Anatomic changes after endo-vascular grafting for aneurysmal disease. Semin Vasc Surg1999;12:192-198.

162 Pages et al. Annals of Vascular Surgery