VASCULAR–ARTERIALMRA Diagnosis

MR Angiography of the Vascular TreeFrom the Aorta to the Foot: Combin-ing Two-dimensional Time-of-flightand Three-dimensional Contrast-en-hanced Imaging. Wei Li, MingZhang, Scott Sher, et al. J Magn ResonImaging 2000; 12:884–889. (W.L., MRIResearch, Department of Radiology,Evanston Northwestern Healthcare,2650 Ridge Avenue, Evanston, IL60201)

• A composite approach for magneticresonance (MR) angiography of the lowerextremities is described. Thirty patientswere studied with this approach, whichcombined a two-dimensional (2D) time-of-flight (TOF) technique with a 3D con-trast-enhanced technique. A head/neckcoil was selected for imaging mid-foot toupper calf, and the body coil was used forthe remainder of the peripheral vasculartree. Acquired data were transferred to aworkstation for postprocessing. The finalmaximum intensity projection images,which display the entire vascular anatomyfrom aortic bifurcation to foot, were cre-ated using a 1024 3 1024 matrix. Verysmall arteries can be differentiated in crit-ical regions like the calf and foot. Com-pared with TOF-2D alone, the scan timewas reduced. This method offers anotheroption for MR angiography of the lowerextremities.Authors’ Abstract

Contrast-enhanced Breath-Hold Three-dimensional Magnetic Resonance An-giography in the Evaluation of RenalArteries: Optimization of Techniqueand Pitfalls. Anil N. Shetty, KostakiG. Bis, Matthias Kirsch, et al. J MagnReson Imaging 2000; 12:912–923. (A.N.S.,Department of Diagnostic Radiology,William Beaumont Hospital, 3601 West13 Mile Road, Royal Oak, MI 48073)

• The authors describe the optimiza-tion of a contrast-enhanced, breath-held,three-dimensional magnetic resonanceangiography (CE-BH-3DMRA) tech-nique in the assessment of the renal ar-teries and compare its utility with con-ventional x-ray angiography (XRA).Signal optimization using specific pulsesequence parameters was based on thepatient’s circulatory conditions, injectionrate, and pulse sequence timing. Fifty-one patients (27 M, 24 F; mean age 69.7years) were evaluated with CE-BH-

3DMRA and XRA. All patients had anMR angiogram 3 months either before orafter XRA. A test bolus study was per-formed for accurate assessment of transittime in each patient. A total of 51 pa-tients (115 vessels) were studied inwhich the sensitivity and specificity forall renal artery stenoses including theproximal and mid-renal arterial seg-ments were 96% and 92%, respectively.In-stent stenosis could only be diag-nosed by quantifying flow beyond thestent using an additional triggered phasecontrast cine pulse sequence. A total of11 accessory renal arteries were correctlyidentified. In addition, fibromusculardysplasia in two patients and stents inthree patients were correctly identifiedon MRA.Authors’ Abstract

Magnetic Resonance Angiography Isan Accurate Imaging Adjunct to Du-plex Ultrasound Scan in Patient Selec-tion for Carotid Endarterectomy.Martin R. Back, Jeffrey S. Wilson, Greg-ory Rushing, et al. J Vasc Surg 2000;32:429–40. (M.R.B., Division of Vascu-lar Surgery, University of South FloridaCollege of Medicine, Harbourside Med-ical Tower #650, 4 Columbia Dr,Tampa, FL 33606)

• Purpose: The purpose of this studywas to evaluate the accuracy of magneticresonance angiography (MRA) for cate-gorizing the severity of carotid diseaserelative to duplex ultrasound scan andcerebral contrast arteriography (CA) todetermine of MRA imaging could re-place the need for cerebral angiographyin cases of indeterminate or inadequateduplex scan imaging. Methods: Seventy-four carotid bifurcations in 40 patientsundergoing 45 carotid endarterectomiesfrom 1996 to 1998 were imaged with du-plex ultrasound scan; MRA (two-dimen-sional neck and three-dimensional intra-cranial, time-of-flight technique); andbiplanar, digital subtraction cerebralarteriography. Studies were blindly re-viewed by one reader who used estab-lished threshold velocity criteria for theduplex scan and the North AmericanSymptomatic Carotid EndarterectomyTrial method for MRA and CA to deter-mine the percentage of diameter reduc-tion of the internal carotid artery (ICA).Disease severity was grouped into fourcategories (,50%, 50%–74%, 75%–99%stenosis and occlusion), and the resultsof MRA and duplex ultrasound scan

were compared with CA. Results: Sensi-tivity, specificity, positive predictivevalue, and negative predictive value fordetection of .50% ICA stenosis were100%, 96%, 98%, and 100% for MRA and100%, 72%, 88%, and 100% for duplexultrasound scan, respectively; similarly,for detection of .75% ICA stenosis val-ues were 100%, 77%, 76%, and 100% forMRA and 90%, 74%, 72%, and 91% forduplex ultrasound scan, respectively.Both MRA and duplex ultrasound scanaccurately differentiated all cases of.95% stenosis (n 5 7) from occlusion (n5 4). Short length ICA flow gaps werepresent on MRA in all cases of 75% to99% stenosis and one half of cases ofCA-defined 50% to 74% stenosis. In pa-tients with 50% to 74% stenosis, themean angiographic stenosis was signifi-cantly greater when a flow gap waspresent on MRA (64% 6 6%) versus noflow gap (57% 6 7%) (P 5 .04). Therewas overall agreement among duplex ul-trasound scan, MRA, and CA in 73% ofcarotids imaged. Of the 24% discordantresults between MRA and duplex ultra-sound scan, MRA correctly predicteddisease severity in all cases, and inaccu-rate duplex ultrasound scan results weredue to overestimation in 83% of cases.The operative plan was altered by CAfindings in only one patient (2%) afterduplex ultrasound scan and MRA. Con-clusion: MRA can accurately categorizethe severity of carotid occlusive disease.Duplex ultrasound scan facilitates pa-tient selection for carotid endarterec-tomy in most cases, but adjunct use ofMRA improves diagnostic accuracy for.75% stenoses and may obviate theneed for cerebral arteriography whenduplex scan results are inconclusive ordemonstrate borderline disease severity.Authors’ Abstract

MR Interventions

Selective Contrast-enhanced MR An-giography. Clemens Bos, Henk F.M.Smits, Chris J.G. Bakker, et al. MagnReson Med 2000; 44:575–582. (C.B.,AZU, Room E01.334, Heidelberglaan100, 3584 CX Utrecht, The Netherlands)

• In this study the feasibility of intraar-terial contrast administration was investi-gated. Its use for navigation and treatmentevaluation during MR-guided intravascu-lar interventions was explored in phantomand animal experiments. An injection pro-tocol was developed, which accounts for

Abstracts of Current Literature

669

sequence parameters and vessel flow rate.Tracking a bolus of contrast agent wasuseful to verify the catheter tip positionand to assess flow conditions. Comparedto intravenous contrast-enhanced mag-netic resonance angiography (CE-MRA),selective contrast administration permit-ted a strongly reduced dose. In two-di-mensional (2D) acquisitions overlap ofvessels was prevented. Injection and ac-quisition were easily and accurately syn-chronized in selective 3D CE-MRA, and ahigh contrast concentration could bemaintained during the entire acquisition.Selective injection is useful in the course ofan intervention, to facilitate navigation,provide information on flow conditions,and to evaluate treatment progressrepeatedly.Authors’ Abstract

Heating Around Intravascular Guide-wires by Resonating RF Waves.Maurits K. Konings, Lambertus W. Bar-tels, Henk F.M. Smits, et al. J Magn Re-son Imaging 2000; 12:79–85. (M.K.K.,Department of Radiology, Image Sci-ences Institute, Room AZU-E.01,334,University Hospital Utrecht, Heidelber-glaan 100, 3584 CX Utrecht, The Neth-erlands)

• We examined the unwanted radio-frequency (RF) heating of an endovascu-lar guidewire frequently used in inter-ventional magnetic resonance imaging(MRI). A Terumo guidewire was partlyimmersed in an oblong saline bath tosimulate an endovascular intervention.The temperature rise of the guidewiretip during an FFE sequence [averagespecific absorption rate (SAR) 5 3.9W/kg] was measured with a Luxtronfluoroscopic fiber. Starting from 26°C,the guidewire tip reached temperaturesup to 74°C after 30 seconds of scanning.Touching the guidewire may cause sud-den heating at the point of contact,which in one instance caused a skinburn. The excessive heating of a linearconductor like the guidewire can only beexplained by resonating RF waves. Thecapricious dependencies of this reso-nance phenomenon on environmentalfactors have severe consequences forpredictability and safety guidelines.Authors’ Abstract

Stents

Predicting Outcome of Angioplastyand Selective Stenting of Multiseg-ment Iliac Artery Occlusive Disease.Richard J. Powell, Mark Fillinger,Daniel B. Walsh, et al. J Vasc Surg 2000;32:564–569. (R.J.P., Section of VascularSurgery, Dartmouth-Hitchcock MedicalCenter, One Medical Center Drive, Leb-anon, NH 03756)

• Background: Patients who require an-gioplasty and stenting of multiple iliac ar-terial segments often require reinterven-tion to maintain long-term patency.Morphologic predictors and causes of fail-ure are unknown. The purpose of the cur-rent study was to define arteriographicpredictors of angioplasty and selectivestent failure in the treatment of multiseg-ment iliac occlusive disease. Methods: Alliliac segments (two common and two ex-ternal) of 75 patients who underwent an-gioplasty and selective stent placement formultisegment iliac occlusive disease (twosegments) were scored through use of amodification of the Society of Cardiovas-cular and Interventional Radiology classi-fication for iliac angioplasty (0 5 no lesion;4 5 most severe). Total iliac score wascalculated by summing scores from eachsegment. A separate external iliac scorewas calculated by adding only the externaliliac scores. Arteriograms were reviewedinitially and at the time of lesion recur-rence and stratified by lesion location andprevious intervention. Results: The area ofprevious endovascular intervention wasthe site of recurrence in 75% of patients.New lesions, presumably a result of pro-gressive atherosclerosis, occurred in 15%of patients, and lesions occurred in bothnew and previously treated iliac segmentsin 10% of patients. Only the external iliacscore was an independent predictor offailed endovascular therapy despitereintervention. For patients with an exter-nal iliac score of 2 or less, the endovascularprimary-assisted patency rates at 6, 12,and 24 months were 96%, 92%, and 89%,respectively. This was improved in com-parison with the 90%, 63%, and 45% pa-tency rates observed in patients with anexternal iliac score of 3 or more (P 5 .001).Patients with an external iliac score of 3 ormore had a significantly lower incidenceof hemodynamic and clinical improve-ment after intervention and a threefoldhigher need for surgical inflow proceduresthan patients with an external iliac score of2 or less. Conclusions: Lesion formation af-ter treatment of multisegment iliac occlu-sive disease typically occurs in areas ofprior intervention. The extent of externaliliac disease can be used to stratify patientswith multisegment iliac occlusive diseasewho will likely respond to endovasculartreatment with a durable result. Patientswith extensive external iliac disease (score.3) have poor results after angioplastyand selective stenting as applied in thisstudy, even with endovascular reinter-vention. They are ideal subjects for pro-spective comparative studies of competinginitial therapies, including stenting, en-dografting, and aortobifemoral bypassgrafting.Authors’ Abstract

Endovascular Stent-Grafts

Adverse Consequences of Internal Il-iac Artery Occlusion during Endovas-cular Repair of Abdominal Aortic An-eurysms. Laura A. Karch, Kim J.Hodgson, Mark A. Mattos, et al. J VascSurg 2000; 32:676–683. (K.J.H., Chief,Section of Peripheral Vascular Surgery,PO Box 19638, Springfield, IL 62794–9638)

• Objective: Embolization of the inter-nal iliac artery (IIA) may be performedduring endovascular abdominal aortic an-eurysm (AAA) repair if aneurysmal dis-ease of the common iliac artery precludesgraft placement proximal to the IIA orifice.The IIA may also be unintentionally oc-cluded because of iliac trauma or coverageby the endograft. The purpose of thisstudy was to determine the incidence, eti-ology, and consequences of IIA occlusionduring endoluminal AAA repair. Methods:Over 2 years, 96 patients have undergoneendoluminal AAA repair. The details ofthe operative procedure, reasons for IIAocclusion, perioperative complications,and clinical follow-up were recorded. Re-sults: The IIA was intentionally occludedin 15 patients (16%) to treat 13 commoniliac artery aneurysms, one IIA aneurysm,and one external iliac artery aneurysm.The IIA was unintentionally occluded in 9patients (9%), resulting from traumatic il-iac dissection in 5 patients and coverage ofthe IIA by the endograft in the remaining4 patients. Three patients had colonischemia. One patient with a unilateral IIAocclusion had sigmoid infarction necessi-tating resection. The other two patientsunderwent intentional occlusion of oneIIA followed by unintentional occlusion ofthe contralateral IIA because of a trau-matic iliac dissection. Both had postoper-ative abdominal pain and distention;rectosigmoid ischemia was revealedthrough colonoscopy. Conservative treat-ment with bowel rest and broad-spectrumantibiotics was successful in both cases.Nondisabling hip and buttock claudica-tion occurred in seven patients (32%) at 1month but resolved by 6 months in threeof these patients. Conclusion: Embolizationof the IIA for iliac aneurysmal disease andunintentional IIA occlusion due to traumaor graft coverage occurs in a considerablenumber of patients undergoing endolumi-nal AAA repair. Most patients with uni-lateral occlusion do not experience colonischemia or disabling claudication. There-fore, unilateral embolization of the IIA iswell tolerated and allows for the endolu-minal treatment of patients with both anAAA and an iliac artery aneurysm,thereby expanding the number of patientswho can be managed with an endovascu-lar approach. Although acute, bilateral IIAocclusions should be avoided, significant

670 • Abstracts May 2001 JVIR

consequences were not observed in oursmall series of patients.Authors’ Abstract

Safety of Coil Embolization of the In-ternal Iliac Artery in EndovascularGrafting of Abdominal Aortic Aneu-rysms. Frank J. Criado, Eric P. Wil-son, Omaida C. Velazquez, et al. J VascSurg 2000; 32:684–688. (F.J.C., 3333North Calvert Street, Suite 570, Balti-more, MD 21218)

• Purpose: During endovasculargrafting of an abdominal aortic aneu-rysm (AAA), iliac limb extension to theexternal iliac artery may be indicatedwhen the common iliac artery is ectaticor aneurysmal. Preliminary or concomi-tant coil embolization of the internal iliacartery (IIA) is thus necessary to preventpotential reflux and endoleak. Wesought to determine the safety of hypo-gastric flow interruption in this setting.Methods: We retrospectively reviewed156 patients who underwent stent-graftAAA repair at two institutions betweenFebruary 1, 1998, and January 31, 1999.Coil embolization of one or both IIAswas undertaken when the diameter ofthe common iliac artery was more than20 mm to enable limb endograft exten-sion to the external iliac artery. Bilateralprocedures were staged. Results: Thirty-nine (25%) of 156 patients were selectedfor coil embolization of one (n 5 28) orboth (n 5 11) IIAs. The interventionswere performed before (n 5 31) or dur-ing (n 5 8) the stent-graft procedure.Complications included groin hemato-mas in 3 patients, iliac artery dissectionin 1, failure to catheterize the IIA in 2,and transient rise in the serum creatininelevel in 3. One patient had erectile dys-function, and five patients (13%) hadbuttock claudication after unilateralocclusion. Serious ischemic complica-tions were not observed. Conclusion: Coilembolization of one or both IIAs appearsto be safe in the setting of endovasculargrafting of AAA. Buttock claudication isa relatively significant problem and maylimit applicability of this strategy to pa-tients who are unfit for standard openrepair.Authors’ Abstract

Relationship between PreoperativePatency of the Inferior Mesenteric Ar-tery and Subsequent Occurrence ofType II Endoleak in Patients Under-going Endovascular Repair of Abdom-inal Aortic Aneurysms. Omaida C.Velazquez, Richard A. Baum, Jeffrey P.Carpenter, et al. J Vasc Surg 2000; 32:777–788. (Ronald M. Fairman, Depart-ment of Surgery, 4 Silverstein Pavilion,3400 Spruce Street, Hospital of the Uni-

versity of Pennsylvania, Philadelphia,PA 19104)

• Objectives: The purpose of thisstudy was (i) to find out whether preop-erative inferior mesenteric artery (IMA)patency (on radiographic imaging) pre-dicts IMA-related endoleaks after endo-vascular repair of infrarenal abdominalaortic aneurysms, (ii) to determine feasi-bility of measuring aneurysm sac pres-sures in patients with endoleaks, and(iii) to report early evidence of effectiveendovascular obliteration of IMAendoleaks. Methods: We studied 76 con-secutive cases of infrarenal aortic aneu-rysms that were repaired with an endo-vascular approach (March 1998–April1999). Results: There were 13 (17%) en-doleaks persistent 30 days after theprocedure. Eleven (85%) of these 13 wereIMA-related endoleaks, which were doc-umented with selective superior mesen-teric artery angiography. The preopera-tive finding (on computed tomographicscan) of a patent IMA does not alwayspredict an IMA-related endoleak, but re-sults in a statistically and clinically sig-nificant higher ratio of patients withIMA-related endoleaks in the immediatepostoperative period (24% versus 3%, P, .035). In eight of the 11 patients withpersistent IMA-related endoleaks, mea-surement of intra-aneurysm sac pres-sures was possible, and six of these pa-tients had systemic pressures within theexcluded aneurysm sac. Nine (82%) of 11IMA-related endoleaks were success-fully obliterated by means of selectiveIMA embolization. Conclusions: Manyendoleaks are caused by a patent IMA,and this can result in persistence of sys-temic pressure within the aneurysm sac.The preoperative finding (on computedtomographic scan) of a patent IMA is apredictor of increased rates of IMA en-doleaks, and IMA endoleaks can be success-fully obliterated through endovascular pro-cedures, after endovascular abdominalaortic aneurysm repair.Authors’ Abstract

Transrenal Fixation of Aortic Stent-Grafts for the Treatment of InfrarenalAortic Aneurysmal Disease. Paul G.Bove, Graham W. Long, Gerald B.Zelenock, et al. J Vasc Surg 2000; 32:697–703. (P.G.B., William BeaumontHospital, 102-VSC, 3601 W. ThirteenMile Rd, Royal Oak, MI 48073)

• Purpose: We evaluated our early ex-perience with the transrenal fixation ofaortic stent-grafts to determine the effi-cacy of this procedure and its effects onrenal artery patency and hemodynamics.Methods: Twenty-eight patients (22 men)had endoluminally placed modular bi-furcated stent-grafts with a bare springstructure at the proximal end crossing

the origin of both renal arteries; no pa-tient with infrarenal fixation was in-cluded for analysis. The mean age of thepatients was 75 6 7 years (range, 58–86years); the mean aneurysm size was 5.86 0.8 cm (range, 4.7–7.2 cm). Eight pa-tients had preoperative or intraoperativeangiographic evidence of renal arteryatherosclerotic disease, but only fourvessels had luminal narrowing of 50% orgreater. No complications were notedduring stent-graft placement, and all pa-tients have returned for follow-up visits,ranging from 1 to 12 months (mean fol-low-up, 6 6 4 months). Follow-up eval-uations included clinical assessment, du-plex ultrasound scan of the renal arteriesand kidneys, and computed tomo-graphic angiography. Results: No evi-dence of lobular or sublobular perfusiondefects of the renal parenchyma was de-tected postoperatively. Two patients ex-hibited postoperative changes in renalartery hemodynamics—one progressingfrom a 30% diameter reduction to agreater than 60% diameter stenosis at the12-month follow-up visit and one with anormal renal artery preoperatively hav-ing elevated flow velocities indicative ofa greater than 60% stenosis at the1-month visit. Of 19 patients with nor-mal preoperative renal function, onlyone has had persistently elevated serumcreatinine levels. Conclusion: We con-clude from this experience that the trans-renal placement of open stents is safeand effectively excludes the aneurysm,potentially expanding the availability ofthis technique to more patients with ashort infrarenal aortic neck. Long-termfollow-up is essential to determine theoverall efficacy of this technique and toidentify potential effects on renal arteryhemodynamics or kidney function.Authors’ Abstract

Duplex Ultrasound Scanning versusComputed Tomographic Angiographyfor Postoperative Evaluation of Endo-vascular Abdominal Aortic AneurysmRepair. Yehuda G. Wolf, Bonnie L.Johnson, Bradley B. Hill, et al. J VascSurg 2000; 32:1142–1148. (C.K.Z., Divi-sion of Vascular Surgery, Stanford Uni-versity Medical Center, 300 PasteurDrive, H-3642, Stanford, California,94305-564).

• Objective: The purpose of this studywas to compare duplex ultrasound scan-ning and computed tomographic (CT)angiography for postoperative imagingand surveillance after endovascular re-pair of abdominal aortic aneurysm(AAA). Methods: One hundred consecu-tive patients with AAA underwent en-dovascular (Medtronic AneuRx, stentgraft) aneurysm repair and were imagedwith both CT angiography and duplex

Abstracts • 671Volume 12 Number 5

ultrasound scanning at regular intervalsafter the procedure. Each imaging mo-dality was evaluated for technical ade-quacy and for documentation of aneu-rysm size, endoleak, and graft patency.In concurrent scan pairs, accuracy of du-plex scanning was compared with CT.Result: A total of 268 CT scans and 214duplex scans were obtained at intervalsof 1 to 30 months after endovascular an-eurysm repair (mean follow-up interval,9 6 7 months). All CT scans were tech-nically adequate, and 198 (93%) of 214duplex scans were technically adequatefor the determination of aneurysm size,presence of endoleak, and graft patency.Concurrent (within 7 days of each other)scan pairs were obtained in 166 instancesin 76 patients (1–6 per patient). The max-imal transverse aneurysm sac diametermeasured with both methods correlatedclosely (r 5 0.93; P , 0.001) without asignificant difference on paired analysis.In 92% of scans, measurements werewithin 5 mm of each other. Diagnosis ofendoleak on both examinations corre-lated closely (P , .001), and comparedwith CT, duplex scanning had a sensitiv-ity of 81%, a specificity of 95%, a positivepredictive value of 94%, and a negativepredictive value of 90%. Discordant re-sults occurred in 8% of examinations,and in none of these was the endoleakclose to the attachment sites or associ-ated with aneurysm expansion. An en-doleak was demonstrated on both testsin all eight patients who had an en-doleak judged severe enough to warrantarteriography. Graft patency was docu-mented in each instance, without dis-crepancy, with both modalities. Conclu-sions: High-quality duplex ultrasoundscanning is comparable to CT angiogra-phy for the assessment of aneurysm size,endoleak, and graft patency after endo-vascular exclusion of AAA.Authors’ Abstract

Does the Endovascular Repair of Aor-toiliac Aneurysms Pose a RadiationSafety Hazard to Vascular Surgeons?Evan C. Lipsitz, Frank J. Veith, TakaoOhki, et al. J Vasc Surg 2000; 32:704–710. (E.C.L., Assistant Professor of Sur-gery, Montefiore Medical Center, Divi-sion of Vascular Surgery-Vascular Lab-oratory, 111 East 210th Street, Bronx,NY 10467)

• Objectives: Endovascular aortoiliacaneurysm (EAIA) repair uses substantialfluoroscopic guidance that requires con-siderable radiation exposure. Doseswere determined for a team of three vas-cular surgeons performing 47 consecu-tive EAIA repairs over a 1-year period todetermine whether this exposure consti-tutes a radiation hazard. Methods: Twen-ty-nine surgeon-made aortounifemoral

devices and 18 bifurcated devices wereused. Three surgeons wore dosimeters(1) on the waist, under a lead apron; (2)on the waist, outside a lead apron; (3) onthe collar; and (4) on the left ring finger.Dosimeters were also placed around theoperating table and room to evaluate thepatient, other personnel, and ambientdoses. Exposures were compared withstandards of the International Commis-sion on Radiological Protection (ICRP).Results: Total fluoroscopy time was 30.9hours (1852 minutes; mean, 39.4 minutesper case). Yearly total effective bodydoses for all surgeons (under lead) werebelow the 20 mSv/y occupational expo-sure limit of the ICRP. Outside leaddoses for two surgeons approximatedrecommended limits. Lead aprons atten-uated 85% to 91% of the dose. Ring dosesand calculated eye doses were within theICRP exposure limits. Patient skin dosesaveraged 360 mSv per case (range, 120–860 mSv). The ambient (.3 m from thesource) operating room dose was 1.06mSv/y. Conclusions: Although the totaleffective body doses under lead fellwithin established ICRP occupationalexposure limits, they are not negligible.Because radiation exposure is cumula-tive and endovascular procedures arebecoming more common, individualsperforming these procedures must care-fully monitor their exposure. Our resultsindicate that a team of surgeons can per-form 386 hours of fluoroscopy per yearor 587 EAIA repairs per year and remainwithin occupational exposure limits. In-dividuals who perform these proceduresshould actively monitor their effectivedoses and educate personnel in methodsfor reducing exposure.Authors’ Abstract

Incidence and Risk Factors of LateRupture, Conversion, and Death afterEndovascular Repair of InfrarenalAortic Aneurysms: The EUROSTARExperience. Peter L. Harris, S. RaoVallabhaneni, Pascal Desgranges, et al,for the EUROSTAR Collaborators. JVasc Surg 2000; 32:739–749. (P.L.H., Re-gional Vascular Unit, Link 8C, RoyalLiverpool University Hospital, PrescotStreet, Liverpool, L7 8XP, UK)

• Objective: The EUROSTAR (Euro-pean Collaborators on Stent/graft tech-niques for aortic aneurysm repair) Reg-istry was established in 1996 to collectdata on the outcome of treatment of pa-tients with infrarenal aortic aneurysmswith endovascular repair. To date, 88European centers of vascular surgeryhave contributed. The purpose of thestudy was to evaluate the results of thistreatment in the medium term (up to 4years) according to the analysis of“hard” or primary end points of rupture,

late conversion, and death. Patients andMethods: Patients with aortic aneurysmssuitable for endovascular aneurysm re-pair were notified to the EUROSTARData Registry Centre before treatment toeliminate bias due to selective reporting.The following information was collectedon all patients: (i) demographic detailsand the anatomic characteristics of theiraneurysms, (ii) details of the endovascu-lar device used, (iii) complications en-countered during the procedure and theimmediate outcome, (iv) results of con-trast enhanced computed tomographicimaging at 3, 6, 12, and 18 months afteroperation and at yearly intervals there-after, and (v) all adverse events. Life ta-ble analysis was performed to determinethe cumulative rates of (i) death from allcauses, (ii) rupture, and (iii) late conver-sion to open repair. Risk factors for rup-ture and late conversion were identifiedthrough regression analysis. Results: ByMarch 2000, 2464 patients had been reg-istered, and their mean duration of fol-low-up was 12.19 months (SD, 12.3months). There were 14 patients withconfirmed rupture of their aneurysms.The cumulative rate (risk) of rupturewas approximately 1% per year. Emer-gency surgery was undertaken in 12(86%) patients, of whom five (41.6%)suved. Two patients who were nottreated surgically also died, which re-sulted in an overall death rate of 64.5%(9/14) of the patients. Significant riskfactors for rupture were proximal type Iendoleak (P 5 .001), midgraft (type III)endoleak (P 5 .001), graft migration (P 5.001), and postoperative kinking of theendograft (P 5 .001). Forty-one patientsunderwent late conversion to open re-pair with a perioperative mortality rateof 24.4% (10/41). The cumulative rate(risk) of late conversion was approxi-mately 2.1% per year. Risk factors (indi-cations) for late conversion were proxi-mal type I endoleak (P 5 .003), graftmigration (P 5 .001), graft kinking (P 5.001), and distal type I endoleak (P 5.001). Conclusions: Endovascular repairof infrarenal aortic aneurysms with thefirst- and second-generation devices thatpredominated in this study was associ-ated with a risk of late failure, accordingto an analysis of observed hard endpoints of 3% per year. Action taken toaddress the risk factors identified by thestudy may improve results in the future.Authors’ Abstract

Correlation of Hemodynamic Impactand Morphologic Degree of Renal Ar-tery Stenosis in a Canine Model. Ste-fan O. Schoenberg, Michael Rock,Friedrich Kallinowski, et al. J Am SocNephrol 2000; 11:2190–2198. (ArminJust, Department of Cell and Molecular

672 • Abstracts May 2001 JVIR

Physiology, Medical Science ResearchBuilding CB#7545, University of NorthCarolina at Chapel Hill, Chapel Hill,NC 27599)

• In a noninvasive comprehensivemagnetic resonance (MR) examination,the morphologic degree of renal arterystenosis was correlated to correspondingchanges in renal artery flow dynamics.Different degrees of stenosis were cre-ated with the use of a chronically im-planted inflatable arterial cuff in sevendogs. For each degree of stenosis, an ul-trafast three-dimensional gadoliniumMR angiography with high spatial reso-lution was performed, followed by car-diac-gated MR flow measurements withhigh temporal resolution for determina-tion of pulsatile flow profiles and meanflow. Flow was also measured by achronically implanted flow probe. Inthree of the dogs, trans-stenotic pressuregradients (‚P) also were measured viaimplanted catheters. Five different de-grees of stenosis could be differentiatedin the MR angiograms (0%, 30%, 50%,80%, 90%). The MR flow data agreedwith the flow probe within 6 20%. Ste-noses between 30 and 80% gradually re-duced the early systolic peak (Max1) ofthe flow profile but only minimally af-fected the midsystolic peak (Max2) ormean flow. Stenoses of more than 90%significantly depressed mean flow bymore than 50%. The ratio between Max1and Max2 (Rmax1/2) gradually fell withthe degree of stenosis. The onset of sig-nificant mean flow reduction and ?P wasindicated by a drop of Rmax1/2 below1–1.2. Thus, the analysis of high-resolu-tion flow profiles allows detection ofearly hemodynamic changes even at de-grees of stenoses not associated with areduction of mean flow. Rmax1/2 allowsdifferentiation of the grade of hemody-namic compromise for a given morphologicstenosis independent of mean flow in a sin-gle comprehensive MR examination.Authors’ Abstract

Renovascular Disease

Cholesterol Crystal Embolism: A Rec-ognizable Cause of Renal Disease.Francesco Scolari, Regina Tardanico,Roberta Zani, et al. Am J Kidney Dis2000; 36:1089–1109. (F.S., Division ofNephrology, Spedali Civili, P le SpedaliCivili, 1, 25125 Brescia, Italy)

• Cholesterol crystal embolism, some-times separately designated atheroem-bolism, is an increasing and still underdiag-nosed cause of renal dysfunction antemor-tem in elderly patients. Renal cholesterolcrystal embolization, also known as athero-embolic renal disease, is caused by showersof cholesterol crystals from an atheroscle-

rotic aorta that occlude small renal arteries.Although cholesterol crystal embolizationcan occur spontaneously, it is increasinglyrecognized as an latrogenic complicationfrom an invasive vascular procedure, suchas manipulation of the aorta during angiog-raphy or vascular surgery, and after antico-agulant and fibrinolytic therapy. Cholesterolcrystal embolism may give rise to differentdegrees of renal impairment. Some patientsshow only a moderate loss of renal function;in others, severe renal failure requiring dial-ysis ensues. An acute scenario with abruptand sudden onset of renal failure may beobserved. More frequently, a progressiveloss of renal function occurs over weeks. Athird clinical form of renal atheroemboll hasbeen described, presenting as chronic, stable,and asymptomatic renal insufficiency. Therenal outcome may be variable; some pa-tients deteriorate or remain on dialysis,some improve, and some remain withchronic renal impairment. In addition to thekidneys, atheroembolization may involvethe skin, gastrointestinal system, and centralnervous system. Renal atheroembolic dis-ease is a difficult and controversial diagnosisfor the protean extrarenal manifestations ofthe disease. In the past, the diagnosis wasoften made postmortem. However, in thelast decade, awareness of atheroembolic re-nal disease has improved, enabling us tomake a correct premortem diagnosis in anumber of patients. Correct diagnosis re-quires the clinician to be alert to thepossibility. The typical patient is a whiteman aged older than 60 years with a base-line history of hypertension, smoking, andarterial disease. The presence of a classictriad characterized by a precipitating event,acute or subacute renal failure, and periph-eral cholesterol crystal embolizationstrongly suggests the diagnosis. The confir-matory diagnosis can be made by means ofbiopsy of the target organs, including kid-neys, skin, and the gastrointestinal system.Thus, Cinderella and her shoe now can bewell matched during life. Patients with renalatheroemboli have a dismal outlook. A spe-cific treatment is lacking. However, it is animportant diagnosis to make because it maysave the patient from inappropriatetreatment. Finally, recent data suggest thatan aggressive therapeutic approach with pa-tient-tailored supportive measures may beassociated with a favorable clinical outcome.Authors’ Abstract

Carotid

Initial Evaluation of Carotid Angio-plasty and Stenting with Three Differ-ent Cerebral Protection Devices. JuanC. Parodi, Ricardo La Mura, L. Mari-ano Ferreira, et al. J Vasc Surg 2000;32:1127–1136. (J.C.P., ICBA, Blanco En-calada 1543/47, Ciudad de BuenosAires 1428, Argentina)

• Objective: The purpose of the studywas to assess the effectiveness of cerebralprotection devices during carotid artery an-gioplasty and stent placement. Methods: Be-tween September 1998 and September 1999,carotid angioplasty and stenting were per-formed in 46 patients with symptomatic(39.1%) or asymptomatic (60.9%) severe ca-rotid artery stenosis. Wallstents were used inall patients with selective predilation. Cere-bral protection devices were used in 25 ofthese patients. Primary end points wereperioperative neurologic complications andmortality. Data were collected prospectively.Results: The overall combined end point ofall neurologic deficits and death rate was4.34%. Two neurologic events (one transientischemic attack and one minor stroke) oc-curred in the unprotected group (9.53%) ver-sus none in the group with cerebralprotection. This difference is not statisticallysignificant. The mortality rate was 0% forboth groups. On an intention to treat basis,the overall technical success rate for carotidangioplasty was 97.8%, and for placement ofcerebral protection devices it was 100%. Animportant number of particles of differentsizes were captured in all cases in whichcerebral protection devices were used. Con-clusion: Experience has shown that cerebralprotection during carotid angioplasty andstenting is technically feasible and appearsto be effective in preventing procedure-re-lated neurologic complications. Further in-vestigation is warranted.Authors’ Abstract

Vena Cava Filters

The Percutaneous Greenfield filter:Outcomes and Practice Patterns.Lazar J. Greenfield and Mary C. Proc-tor. J Vasc Surg 2000; 32:888–893.(L.J.G., University of Michigan Depart-ment of Surgery, 2101 Taubman Cen-ter/Box 0346, Ann Arbor, MI 48109-0346)

• Objective: The percutaneous steelGreenfield filter (PSGF) is similar in ap-pearance to the titanium Greenfield filter(TGF) but differs in the length and ori-entation of the attachment hooks and inthe over-the-wire delivery system. Be-cause these differences improve ease ofinsertion and attachment, they may af-fect patient outcomes and physicianpractices. The purpose of this study wasto evaluate the performance of the PSGFrelative to the TGF and to determinewhether there had been a change in phy-sician practices. Methods: The MichiganFilter Registry contains data for a pro-spective cohort of 2188 patients withGreenfield filters. Procedural and long-term outcomes for patients with a PSGFwere abstracted. These events were com-pared with rates for Registry patientswho had a TGF. Trends for indication for

Abstracts • 673Volume 12 Number 5

placement, delivery route, and filter lo-cation were also compared with pub-lished series. Results: Since 1995, 600PSGFs have been placed in 599 patients.A 1-year mortality rate of 42% left 349patients available for annual follow-up,and studies were completed for 231(66%). Periprocedural events occurred in2.5% of cases with associated morbidityin 1.5%. The rate of new pulmonary em-bolism was 2.6%, and vena caval pa-tency was 98.3%. The combined rate ofnew venous thromboembolic events was12.5%. Left-sided femoral vein place-ments increased to 20%, and the majorindication for filter placement has be-come prophylaxis (46%). Conclusions:The PSGF is similar to the TGF withrespect to patient outcomes, and it pro-vides decreased rates of asymmetryalong with excellent fixation. The flexi-ble carrier system has allowed more fre-quent access through the left femoralvein. The ease of use and favorable pa-tient outcomes have resulted in morefrequent placement for prophylacticindications.Authors’ Abstract

Lessons Learned from a 6-year Clini-cal Experience with Superior VenaCava Greenfield Filters. Enrico As-cher, Anil Hingorani, Boris Tsemekhin,et al. J Vasc Surg 2000; 32:881–887.(E.A., Director, Division of VascularSurgery, Maimonides Medical Center,4802 Tenth Avenue, Brooklyn, NY11219)

• Purpose: Therapy to prevent pulmo-nary embolism (PE) resulting from up-per extremity deep venous thrombosis(UEDVT) remains controversial despitean increasing incidence of DVT of upperextremity origin. The purpose of thisstudy was to evaluate the results of 72superior vena cava Greenfield filters(SVC-GFs) placed in patients at risk forPE arising from UEDVT. Methods: Dur-ing the past 78 months, we placed SVC-GFs in 72 patients with UEDVT in whomanticoagulation was either deemed con-traindicated (n 5 67) or proved ineffec-tive in preventing recurrent PE (n 5 4) orextension of the thrombus (n 5 1). Therewere 25 male (35%) and 47 (65%) femalepatients whose ages ranged from 25 to99 years (mean, 74 years). Follow-upranged from 10 days to 78 months(mean, 7.8 months). Sequential chest ra-diographs revealed no filter migration ordisplacement in 26 patients. Results:Thirty-four patients died in the hospitalof causes unrelated to the SVC filter orrecurrent thromboembolism (mean timeto death, 20 days). Follow-up of thesurving 38 patients ranged from 1 monthto 78 months (mean, 22 months); none ofthese patients were seen with any evi-

dence of PE. One SVC-GF was incor-rectly discharged into the innominatevein and left in place. This vein remainspatent 2 months after insertion withoutevidence of filter migration. Conclusions:We think that insertion of SVC-GFs is asafe, efficacious, and feasible therapyand may prevent recurrent thromboem-bolism in patients with UEDVT who areresistant to anticoagulation or have con-traindications to anticoagulation.Authors’ Abstract

VASCULAR–VENOUSDialysis Access

Access Recirculation in TemporaryHemodialysis Catheters as Measuredby the Saline Dilution Technique.Mark A. Little, Peter J. Conlon, John J.Walshe. Am J Kidney Dis 2000; 36:1135–1139. (J.J.W., Consultant Nephrologist,Beaumont Hospital, Dublin 9, Ireland)

• Ultrasound dilution technology isemerging as the standard for measuringaccess recirculation and blood flow in he-modialysis patients. In temporary dialysiscatheters, studies using the traditionaltwo-needle urea method have suggestedthat short femoral catheters are associatedwith an unacceptably high degree ofrecirculation. This problem has neverbeen assessed using ultrasound dilutiontechnology. We performed a prospectiveobservational study of consecutive pa-tients undergoing dialysis through a tem-porary catheter. Measurements weremade on 49 catheters; 10 catheters wereexcluded because poor flow necessitatedreversal of the dialysis ports. Thirty-ninecatheters in 33 patients were included inthis analysis, of which 26 catheters werelocated in the femoral vein, and 13 cathe-ters, in the internal jugular vein. Dialyzerblood flow was adjusted to give an ultra-sonic flow rate of 250 mL/min (actualmean blood flow, 234.3 mL/min; 95% con-fidence interval [CI], 228 to 241). Overallmean recirculation rate was 8.9% (95% CI,4.8 to 13.0). Multivariate analysis showedcatheter location and length to be indepen-dent predictors of recirculation. Bloodflow (within the range tested), durationinto dialysis, time since catheter insertion,cardiac rhythm, and catheter type had nosignificant effect on recirculation rates. Re-circulation in femoral catheters (13.1%)was significantly greater than that in inter-nal jugular catheters (0.4%; P , 0.001).Femoral catheters shorter than 20 cm hadsignificantly greater recirculation (26.3%)than those longer than 20 cm (8.3%; P 50.007). We conclude that temporary femo-ral catheters shorter than 20 cm are asso-ciated with increased recirculation rates.In addition, when dialysis dose deliveryis a priority, locating the temporary cath-

eter in the internal jugular vein is anadvantage.Authors’ Abstract

Risk of Bacteremia from TemporaryHemodialysis Catheters by Site of In-sertion and Duration of Use: A Pro-spective Study. Matthew J. Oliver,Sandra M. Callery, Kevin E. Thorpe, etal. Kidney Int 2000; 38:2543–2545. (M.O.,Division of Nephrology, Room A2039,Sunnybrook & Women’s CollegeHealth Sciences Centre, Toronto, ON,Canada M4N 3M5)

• Background: Uncuffed, nontunneledhemodialysis catheters remain the pre-ferred means to gain immediate accessto the circulation for hemodialysis. Bac-teremia is the primary complication thatlimits their use. The risk of bacteremiaby site of insertion and duration of usehas not been well studied. Methods: Twohundred eighteen consecutive patientswho required a temporary hemodialysiscatheter were prospectively followed.Results: Catheters were placed at 318new insertion sites and remained in usefor a total of 6235 days. The incidence ofbacteremia was 5.4% after three weeks ofplacement in internal jugular vein and10.7% after one week in femoral vein[relative risk for bacteremia 3.1 (95% CI,1.8 to 5.2)]. The incidence of bacteremiawas 1.9% one day after the onset of anexit site infection but increased to 13.4%by the second day if the catheter was notremoved. Guidewire exchange for mal-function and patient factors did not sig-nificantly affect the risk of bacteremia.Conclusions: Internal jugular cathetersmay be left in place for up to three weekswithout a high risk of bacteremia, butfemoral catheters in bed-bound patientsshould be removed after one week.Catheter exchanges over a guidewire forcatheter malfunction do not increasebacteremia rates. Temporary cathetersshould be removed immediately if anexit site infection occurs.Authors’ Abstract

Tunneled-Cuffed Catheter AssociatedInfections in Hemodialysis PatientsWho Are Seropositive for the HumanImmunodeficiency Virus. Michele H.Mokrzycki, Bernd Schroppel, Gero vonGersdorff, et al. J Am Soc Nephrol 2000;11:2122–2127. (M.H.M., Associate Pro-fessor of Clinical Medicine, 3332 Roch-ambeau Avenue, Centennial 423, Mon-tefiore Medical Center, Bronx, NY10467)

• Infection rates in tunneled-cuffedcatheters (TCC) are reported to be higherin immunocompromised patients. Thepurpose of this study was to evaluateTCC-associated infection rates in pa-tients with HIV infection (HIV1). Data

674 • Abstracts May 2001 JVIR

were collected in 40 HIV1 patients and41 controls (C), and in 118 TCC (HIV1,58; C, 60) for 28,146 catheter days(HIV1, 16,227;C, 11,919). There were nosignificant differences in the TCC bacte-remia rates (HIV1, 2.23 versus C, 2.53per 1000 TCC days, P 5 NS) or in theTCC exit site infection rates (HIV1, 2.20versus C, 2.24 per 1000 TCC days, P 5NS) between the groups. The number ofTCC removed due to infection was alsosimilar, (HIV1 versus C: 17 versus 15%,P 5 NS). In the HIV1 group, the asso-ciation of hepatitis B surface antigen-emia with TCC exit site infection wasdependent on the history of injectiondrug use. Black race was a significantrisk factor for higher TCC exit site infec-tion rates. whereas prophylactic antibi-otic use and high CD4 count were sig-nificantly associated with lower TCCexit site infection rates. None of the fac-tors significantly predicted bacteremiarate in either group (HIV1 or C). In com-parison to controls, HIV1 patients had afivefold increased risk of having a Gram-negative organism (P 5 0.02) and a sev-enfold increased risk of a fungal isolate(P 5 0.08), although the latter findingwas not statistically significant. HIV in-fection is not a significant risk factor forTCC-associated infection but is associ-ated with a higher prevalence of Gram-negative and fungal species.Authors’ Abstract

Treatment of Stenosis and Thrombo-sis in Haemodialysis Fistulas andGrafts by Interventional Radiology.Luc Turmel-Rodrigues, Josette Pen-gloan, Serge Baudin, et al. Nephrol DialTransplant 2000; 15:2029–2036. (L.T.R.,Department of Cardio-Vascular Radiol-ogy, Clinique St-Gatien, 8 Place de laCathedrale, F-37000 Tours, France)

• Background: There are no large se-ries reporting the long-term results afterradiological treatment of both stenosisand thrombosis in native fistulas (AVFs)and prosthetic grafts. Methods: Between1987 and 1999, 726 dilations, 135 stentplacements and 257 declotting proce-dures were performed in 209 consecu-tive forearm AVFs, 74 upper arm AVFsand 156 prosthetic grafts. The stentsused were the Wallstent, the Craggstent,and the Passager. Declotting was per-formed by manual catheter-directedthromboaspiration, with or without pre-vious urokinase infusion. Results: Theinitial success rates ranged from 78 to98%. The rate of significant complica-tions was 2%. Primary patency rates at 1year were twice as good for forearmAVFs (50%) than for grafts (25%) (P ,.05), and were 34% for upper arm AVFs.Secondary patency rates were similar inthe 3 groups at 1 year (80–86%) and at 2

years (68–80%). Reintervention was nec-essary every 18 months in forearm AVFscompared to every 9 months in grafts (P, .05). Thrombosed grafts fared worsethan failing grafts. Accesses of less than1 year’s duration needed more reinter-ventions than older accesses (every 16months versus 30 in forearm AVFs, ev-ery 7 months versus 13 in grafts, P ,.05). Conclusions: The percutaneous treat-ment of stenosis and thrombosis in hae-modialysis access achieves patency ratessimilar to those reported in the surgicalliterature and confirms that grafts mustbe avoided as much as possible giventheir poorer outcome, especially after thefirst thrombosis. Poorer outcome is alsodemonstrated in accesses of less than 1year’s duration.Authors’ Abstract

Improving Longevity of Prosthetic Di-alysis Grafts in Patients with Disad-vantaged Venous Outflow. Alan R.Wladis, Charles L. Mesh, Jean White, etal. J Vasc Surg 2000; 32:997–1005. (C.M.,c/o Cranley Surgical Associates, 310Terrace Avenue, Cincinnati, OH 45220)

• Objective: Angioaccess for hemodi-alysis in an extremity with disadvan-taged venous outflow has reduced long-term patency. We hypothesized thatarteriovenous bridge graft patency couldbe improved in patients with disadvan-taged venous outflow by preoperativevenous duplex mapping. Methods: Thecharts of 114 patients who underwent115 prosthetic arteriovenous bridgegrafts were reviewed. Disadvantagedvenous outflow was defined on the basisof any combination of prior arterio-venous bridge graft, multiple venipunc-tures, and clinical examination. Patientswere grouped according to the presenceor absence of disadvantaged venousoutflow. Three groups were analyzed:those with normal venous outflow whohad an initial arteriovenous bridge graft(NML), those with disadvantaged ve-nous outflow who had only a clinicalexamination before redo arteriovenousbridge graft (REDO/DVO), and thosewith disadvantaged venous outflowwho underwent preoperative duplexscanning venous evaluation (MAP/DVO). Life table primary and secondary12-month patency rates were comparedby means of log-rank analysis. Results:Life table analysis yielded 6-month pri-mary patency rates of 65.9% 6 5.7%,66.4% 6 7.3%, and 43.8% 6 10.9% forNML, MAP/DVO, and REDO/DVO,respectively. The secondary patencyrates at 6 months for NML (91.9% 63.4%) and MAP/DVO (91.1% 6 4.9%)were statistically equivalent, and bothwere significantly better than the pa-tency for REDO/DVO (75.0% 6 10.0%; P

5 .004 and P 5 .04, respectively). Thistrend persisted beyond 12 months. Con-clusion: Preoperative evaluation of ve-nous anatomy in patients with disadvan-taged venous outflow results in anarteriovenous bridge graft patency com-parable to that seen in patients undergo-ing initial arteriovenous bridge grafts.Vein mapping improves arteriovenousbridge graft durability in the patientwith disadvantaged venous outflow byallowing the surgeon to select venousreturn that is in direct continuity withthe central venous system.Authors’ Abstract

Relation Between Gender and Vascu-lar Access Complications in Hemodi-alysis Patients. Brad C. Astor, JosefCoresh, Neil R. Powe, et al. Am J Kid-ney Dis 2000; 36:1126–1134. (J.C., 2024 EMonument St, Ste 2-600, Baltimore, MD21205)

• Native arteriovenous (AV) fistulaefor hemodialysis vascular access are be-lieved to be associated with fewer com-plications than synthetic polytetrafluo-roethylene (PTFE) grafts. We conducteda study among patients in the DialysisMorbidity and Mortality Study to com-pare risk factors for complications of AVfistulae and PTFE grafts in men andwomen and to examine the effect of ageon vascular access complications. We an-alyzed data from 833 incident patientswith end-stage renal disease who had aPTFE graft (n 5 621) or AV fistula (n 5212) in use 1 month after starting hemo-dialysis therapy. Follow-up using inpa-tient and outpatient Medicare adminis-trative data identified a 1.8-times greaterrisk for a subsequent vascular accessprocedure for PTFE grafts (0.71 proce-dures/access-year) than for AV fistulae(0.39 procedures/access-year). Men withgrafts and women with grafts or fistulaehad a greater risk for a first subsequentaccess procedure than did men with fis-tulae (0.79, 0.65, and 0.59 versus 0.33procedures/access-year, respectively).After adjustment for age, race, presenceof diabetes mellitus, and history ofsmoking, peripheral vascular disease,and cardiovascular disease, use of aPTFE graft compared with an AV fistulawas associated with a greater risk for afirst subsequent procedure in men (rela-tive hazard, 2.2; 95% confidence interval[Cl], 1.6 to 2.9), but not in women (rela-tive hazard, 1.0; 95% Cl, 0.7 to 1.4). Theexcess risk associated with a PTFE graftcompared with an AV fistula was lim-ited to men in the lower three quartilesof age (ie, ,72 years). These data raiseconcern that the potential benefits of AVfistulae over PTFE grafts are not realizedin women and older men. A better un-derstanding of the determinants of suc-

Abstracts • 675Volume 12 Number 5

cessful access maturation and mainte-nance in these groups is needed.Authors’ Abstract

TRAUMA

Helical Computed Tomographic Scanin the Evaluation of Mediastinal Gun-shot Wounds. David E. Hanpeter,Demetrios Demetriades, Juan A. Asen-sio, et al. J Trauma 2000; 49:689–695.

• Background: The standard evalua-tion of mediastinal gunshot woundsusually requires angiography and eitheresophagoscopy or esophagography. Inthe present study, we have evaluated therole of helical computed tomographic(CT) scanning in reducing the need forangiographic and esophageal studies.Methods: This was a prospective study ofpatients with mediastinal gunshotwounds who were hemodynamicallystable and would otherwise require an-giography and esophageal evaluation.All patients underwent CT scan of thechest with intravenous contrast to delin-eate the missile trajectory. If the missiletract was in close proximity to the aorta,great vessels, or esophagus, then tradi-tional evaluation with angiographic oresophageal evaluation was pursued. Re-sults: A total of 24 patients met the inclu-sion criteria and underwent CT scanevaluation of their mediastinal gunshotwounds. One patient was taken for ster-notomy to remove a missile embeddedin the myocardium solely on the basis ofthe result of the CT scan. Because ofproximity of the bullet tract, 12 patientsrequired additional evaluation witheight angiograms and nine esophageal

studies. One of these patients had a pos-itive angiogram (bullet resting againstthe ascending aorta) and underwentsternotomy for missile removal; all otherstudies were negative. The remaining 11patients were found to have well-de-fined missile tracts that approached nei-ther the aorta nor the esophagus, and noadditional evaluation was pursued.There were no missed mediastinal inju-ries in this group. Overall, 12 of 24 pa-tients (50%) had a change in manage-ment (either received an operation oravoided additional radiographic or en-doscopic evaluation) on the basis of theCT scan. Conclusion: The helical CT scanprovides a rapid, readily available, non-invasive means to evaluate missiletrajectories. This permits accurate assess-ment of potential mediastinal injury andreduces the need for routine angio-graphic and esophageal studies.Authors’ Abstract

Long-Term Follow-up of Trauma Pa-tients with a Vena Cava Filter. Ran-dolph Wojcik, Mark D. Cipolle, IvyFearen, et al. J Trauma 2000; 49:839–843. (R.W., Department of Surgery, Le-high Valley Hospital, Cedar Crest &I-78, PO Box 689, Allentown, PA 18105-1556)

• Background: Venous thromboembo-lism (VTE) is an important complication inblunt trauma patients. At our Level Itrauma center, we had a deep venousthrombosis (DVT) rate of 3.2% from 1993to 1997 despite an aggressive VTE prophy-laxis program. During this time period, weplaced vena caval filters (VCF) for bothtraditional and prophylactic indications.This project was developed to establish a

VCF registry for trauma patients to deter-mine the long-term complications of VCFplacement. Methods: A letter was sent to alltrauma patients who had a VCF placedfrom 1993 through 1997. Patients wereasked to return for a history and physicalexamination to detect signs and symp-toms related to VTE, a duplex ultrasoundof the inferior vena cava, and a plain ab-dominal radiograph to determine filtermigration. Results: There were 191 VCFsinserted in our trauma population from1993 to 1997. There were 105 patients (75male and 30 female) available for evalua-tion, with a mean follow-up of 28.9months. Forty-one VCFs were placed inpatients with DVT or pulmonary embo-lism, and 64 were placed in patients forprophylactic indications as per the guide-lines developed by the Eastern Associa-tion for the Surgery of Trauma. Therewere no clinically identifiable complica-tions related to insertion of the VCF. Therewere no pulmonary embolisms detectedafter VCF insertion. In follow-up, only onefilter (0.95%) migrated, and this was min-imal (1 cm cephalad). One (0.95%) venacava was occluded, based on duplex ultra-sonography, and 11 patients (10.4%) hadsigns or symptoms of leg swelling a hos-pital discharge. Twenty eight (44% of the64 patients with prophylactic VCFs devel-oped a DVT after filter placement. Conclu-sion: VCFs placed in trauma patients haveacceptable short- and long-term complica-tion rates. Consideration should be givento prophylactic VCF placement in patientsat high risk VTE. Randomized controlledtrials needed to evaluate whether VCF in-sert increases the risk for subsequent DVT.Authors’ Abstract

676 • Abstracts May 2001 JVIR