- Email: [email protected]
Type II endoleaks: Predictable, preventable, and sometimes treatable?
Type II endoleaks: Predictable, preventable, and sometimes treatable? Duncan J. Parry, MBChB, FRCSEd,a David O. Kessel, MA, FRCP, FRCR,b Iain Robertson, FRCP, FRCR,b Lucy Denton, DCR(R),b Jai V. Patel, FRCP, FRCR,b David C. Berridge, DM, FRCSEd, FRCS,a Ralph C. Kester, MD, ChM, FRCS,a and David J. A. Scott, MD, FRCS, FRCSEd,a Leeds, United Kingdom Objective: The purpose of this study was to evaluate the effect of preoperative coil embolization of lumbar and inferior mesenteric arteries on the incidence of type II endoleak after endovascular abdominal aortic aneurysm repair. Methods: The subjects were consecutive patients who underwent EVAR between January 1996 and January 2001. Patent aortic side branches were identified with preprocedural spiral computed tomographic scanning and calibrated angiography. Coil embolization was performed before EVAR. Patients were followed up with plain radiographs and ultrasound and dual phase spiral computed tomographic scans. Digital subtraction angiography was performed when endoleak was suspected. The outcome measures were the incidence of type II endoleaks and changes in maximum aortic sac diameter (Dmax). Results: Forty patients underwent EVAR, with a median duration of follow-up of 24 months (range, 3 to 48 months). Before surgery, the inferior mesenteric artery was patent in 16 patients (45%) and the lumbar arteries in 21 patients (53%). Inferior mesenteric artery embolization was successful in 13 of 16 patients (81%). Lumbar embolization was attempted in 13 patients and was successful in eight (62%). During EVAR, successful sac exclusion was achieved in 38 patients (95%). None of the patients who underwent embolization before EVAR had type II endoleak develop, eight of 13 patients (62%) with patent lumbar arteries had endoleaks develop (P ⴝ .006), and three of these patients subsequently underwent successful coil embolization. Type II endoleak was associated with a 2.0-mm median increase in Dmax (P ⴝ .045). A 3.0-mm median reduction in Dmax was seen in the absence of type II endoleak (P ⴝ .002). Conclusion: Type II endoleaks are predictable, preventable, and sometimes treatable. Significant sac shrinkage occurs in the absence of lumbar endoleak but not in the presence of type II endoleak. (J Vasc Surg 2002;36:105-10.)
Endoleak is a complication specific to endovascular abdominal aortic aneurysm repair (EVAR) and has been defined as blood flow outside the stent graft but within the aneurysm sac.1 Aortic branch vessels (inferior mesenteric artery [IMA] and lumbar arteries) form potential communications between the aneurysm sac and the systemic circulation. Type II branch vessel endoleaks can transmit pressure into the aneurysm sac.2 Malina et al3 noted that even minor endoleak or collateral perfusion inhibits sac shrinkage. More recently, both in vitro and clinical studies have shown pulsatile systemic pressure within the aneurysm sac in the presence of type II endoleak.4,5 Cases of type II endoleak causing both sac growth and rupture also have been reported.6-9 Type II endoleaks have been considered relatively benign,10 but these results indicate their clinical importance. Recent evidence indicates the correlation between side branch patency and type II endoleak.11,12 Preoperative occlusion of these branches may prevent the subsequent development of type II endoleak. From the Departments of Vascular Surgerya and Radiology,b St James’s University Teaching Hospital, Leeds Teaching Hospitals. Competition of interest: nil. Reprint requests: Dr David Kessel, Department of Radiology, St James’s University Hospital, Beckett St, Leeds, LS9 7TF, United Kingdom (email: [email protected]). Copyright © 2002 by The Society for Vascular Surgery and The American Association for Vascular Surgery. 0741-5214/2002/$35.00 ⫹ 0 24/1/125023 doi:10.1067/mva.2002.125023
MATERIALS AND METHODS Design. This study was conducted as a prospective intervention study set in a university teaching hospital. Patient selection. The study population comprised consecutive patients with abdominal aortic aneurysm who underwent EVAR between September 1996 and April 2001. For determination of suitability, patients for EVAR underwent screening with dynamically enhanced spiral computed tomographic (CT) scan and digital subtraction angiography with a calibrated catheter. Patients suitable for EVAR had stent grafts sized according to manufacturer recommendations. All scans were reviewed by at least two vascular radiologists (DK, IR, JP), and patent aortic side branches (IMA and lumbar arteries) were noted. Patency was defined as evidence of perfusion of the vessel from the aortic lumen. For the purpose of the study, we defined a significantly sized lumbar vessel as being amenable to catheterization with a 4F or 5F catheter (ie, measuring approximately ⬎1.5 mm in diameter). Embolization was only attempted in lumbar arteries considered amenable to catheterization (ie, those measuring approximately ⬎1.5 mm in diameter). Preoperative coil embolization of aortic side branches. Preoperative embolization was performed before EVAR using 2-mm to 5-mm coils to occlude the target artery. In the IMA, coils were deployed proximal to the left colic artery (Fig 1). Currently, we attempt embolization in all patients with significantly sized patent lumbar arteries 105
JOURNAL OF VASCULAR SURGERY July 2002
106 Parry et al
lying outside the aneurysm neck. In our early experience, this was not routine, hence lumbar artery embolization was not attempted in all patients. Technical success was defined as angiographic occlusion of the IMA or all target lumbar arteries. Endovascular aortic aneurysm repair. EVAR was performed with general anesthesia in an interventional radiology suite with operating room conditions. Vanguard I (Boston Scientific, St Albans, UK), AneuRx (Medtronic, Watford, UK), and Excluder (Gore, Livingston, UK) stent grafts were used. Graft deployment was performed by a combined team of vascular radiologists and surgeons. Completion angiography was performed to check for endoleaks. Thirty-day morbidity and mortality were documented in all patients. Follow-up assessment. Patients were followed with plain radiographs and ultrasound and dual-phase contrast enhanced CT scans. Dynamically enhanced scans were commenced 30 seconds after the start of contrast infusion, and delayed scans taken 30 seconds later during venous phase enhancement. Scans were reviewed for evidence of patent branch arteries, endoleak, and maximum external diameter of the aneurysm sac. Diameters were measured by hand with mechanical calipers on the hard copy images. The greatest transverse diameter of the aneurysm sac (Dmax) was recorded. When the aneurysm sac was elliptic, the short axis measurement also was recorded. When an apparent increase was found in Dmax, comparison was made between scans to ensure that a comparable level was assessed and a reproducible measurement had been taken. Time of occurrence of type II endoleak was defined as the first instance of demonstrable sac perfusion on CT or ultrasound scan in the absence of a graft-related endoleak. Scans were performed before discharge and at 1, 3, 6, 12, 18, and 24 months and annually thereafter. Results were recorded on a database (DK, LD, DJAS). The site of origin of all leaks was confirmed with angiography, including, if necessary, selective ipsilateral hypogastric artery angiography. Endoleaks were classified as immediate (present from stent graft deployment), early (detected within 30 days of EVAR), or delayed (detected after 1 month during followup). Persistent endoleaks were defined as being detectable on at least two consecutive follow-up scans, at least 3 months apart. Secondary intervention. Embolization was attempted in patients with persistent lumbar endoleaks without shrinkage of the aneurysm sac. The ipsilateral hypogastric artery was catheterized with a 5F catheter, and access to the target lumbar artery was obtained with a coaxial microcatheter (Fig 2). Coils were deployed as close to the aneurysm sac as possible. In most cases, multiple Gianturco coils were deployed in the inflow vessel. If catheterization of the target vessel was impossible, then the feeding vessels were embolized. When collateral vessels perfused the adjacent lumbar arteries, the entire ileolumbar trunk was occluded. Glue (N-butyl 2-cyanoacrylate) embolization was performed in a single patient in whom catheterization of the right fourth lumbar artery, which fed the endoleak, was
Table I. Patent aortic side braches at presentation Patent vessels at presentation IMA All lumbar arteries* IMA and lumbar IMA and/or lumbar
No. 16 (45%) 21 (53%)* 10 (25%) 27 (67.5%)
*Of 21 patients with patent lumbar arteries, 19 (90%) had a patent fourth lumbar artery; 7 (33%) had a patent fourth and second/third lumbar artery; and 2 (10%) had a patent third lumbar artery alone.
impossible. Particulate embolization was not used. Success was defined as abolition of the endoleak on angiography and subsequent imaging studies. Outcome measures. The primary outcome measure was the incidence of type II endoleak. Secondary outcome measures were the technical success rate of preoperative and postoperative embolization and change in the Dmax. Statistical analysis. All data were analyzed with SPSS for Windows version 9.0 (SPSS, Inc, Chicago, Ill). Categoric data were analyzed with a two-tailed Fisher exact test for the difference in proportions of type II endoleak in those cases with successful embolization as compared with those with patent side branches at baseline. Nonparametric data were analyzed with Wilcoxon matched pairs signed rank test for changes in Dmax. RESULTS Forty patients (34 male, six female) underwent treatment with EVAR, with Vanguard I, AneuRx, and Excluder stent grafts deployed in five, 33, and two patients, respectively. Median age was 75.6 years (range, 54 to 92 years). Twenty-seven patients (67.5%) had patent IMA or lumbar arteries (Table I shows the prevalence of patent branch vessels on preoperative imaging). Preoperative embolization. Nineteen of the 27 patients with patent aortic side branches underwent attempted preoperative embolization. The procedures typically took between 30 and 60 minutes and were without complication. Technical success was defined as angiographic occlusion of the IMA or all target lumbar arteries. All successful lumbar vessel embolization involved coil deployment in the larger fourth and third lumbar arteries. The median number of lumbar vessels embolized was two (range, 1 to 5). Embolization was successful in 14 patients (74%), and 21 target vessels (78%) were occluded. Endovascular aneurysm repair. Successful aneurysm exclusion was achieved in 38 of 40 patients (95%). One immediate type III endoleak with no demonstrable outflow vessel was seen. The leak was the result of a tiny “pinholelike” defect in the fabric of the right iliac limb. A single type II endoleak was associated with a solitary patent lumbar artery. Both settled spontaneously by 30 days and did not recur during follow-up. The 30-day mortality rate was four of 40 (10%). Two deaths occurred after resections for colonic malignant disease 2 weeks after EVAR. The need for major abdominal surgery had been an indication for
JOURNAL OF VASCULAR SURGERY Volume 36, Number 1
Parry et al 107
Table II. Patients with lumbar endoleak: origin, changes in sac size, and outcome of secondary intervention
Case 1 2 3 4 5 6 7 8 Median
Source inflow 3 outflow
Time when detected (mo)
⌬ Dmax (mm)
Secondary embolization
Further interventions or events
L L4 3 R L4 L L3 3 R L4 L L4 3 R L4 R L4 3 L L4 R L4 3 L L3/4 L L4 3 R L4 R L4 3 L L4. L L4 3 R L4
18
⫹6
Successful embolization‡
6
0
12
⫹5
6
⫹2
12
0
1
⫹2
12
⫹1
3
⫹5
Failed embolization Awaiting embolization Died before intervention† Failed embolization* Successful embolization Failed embolization Successful embolization Successful embolization
9
⫹2
⌬ Dmax2 –3
0 –7 Died during follow-up† –1 Died during follow-up†
Success rate 3/6 (50%)
*Unable to traverse aneurysm sac. Inflow vessel only embolized. † All deaths were unrelated to stent grafting. ‡ Laproscopic clipping failed, followed by successful second attempt at embolization. L3, third lumbar artery; L4, fourth lumbar artery; R, right; L, left; ⌬ Dmax, change in maximum sac diameter before secondary intervention; ⌬ Dmax2, change in sac diameter after secondary intervention.
EVAR in these patients, and in each case, recovery after the aneurysm repair was uneventful. The third patient died of cholecystitis associated with cardiac failure on the 30th postoperative day. The fourth died of small bowel infarction 48 hours after EVAR. A subsequent postmortem examination revealed extensive atheroma in the suprarenal aorta, and the superior mesenteric artery was patent. Patchy small intestinal infarction was thought to have arisen from massive embolization of atheromatous debris during stent deployment. These latter two patients were octogenarians and were classified as unfit for open aneurysm repair. In addition, both patients had large aneurysms with short markedly angulated proximal aneurysm necks. Both procedures were complex and necessitated placement of additional proximal cuffs. These four patients were all excluded from the subsequent analysis. Follow-up. The median follow-up period for all patients (n ⫽ 36) was 24 months (range, 6 to 48 months), and follow-up periods of the patients with and without embolization were 24 months (range, 6 to 48 months) and 30 months (range, 18 to 36 months), respectively. The incidence rate of lumbar endoleak was eight of 36 (22%) during follow-up (Table II). The median time of lumbar endoleak detection was 9 months (range, 1 to 18 months). All endoleaks persisted on at least two consecutive scans. Seven endoleaks were detected with CT scan, and one with duplex scan. All endoleaks arose from the third or fourth lumbar arteries and had a patent outflow source. Two patients (6%) had a type I endoleak develop (one proximal, one distal), and one patient (3%) a type III endoleak. One patient (3%) with an inflammatory aneurysm had an aor-
toduodenal fistula develop at 6 months that needed stent graft replacement with an in situ silver impregnated graft.13 These patients were all excluded from statistical analysis of changes in sac size. Effect of preoperative embolization on incidence of type II endoleak. No type II endoleaks were seen among patients who had spontaneously occluded aortic side branches at baseline (n ⫽ 13) or who underwent successful IMA or lumbar embolization (n ⫽ 14). No IMA endoleaks were found. Eight of 13 patients (62%) with patent lumbar arteries had a subsequent lumbar endoleak develop, compared with none of the eight patients (0%) who had undergone preoperative lumbar embolization (P ⫽ .006, Fisher exact test). Changes in maximum sac diameter at follow-up. Changes in Dmax are illustrated in Fig 3. In the absence of lumbar endoleak, a mean reduction of 3.0 mm (interquartile range, 10.5-0) was seen in Dmax (P ⫽ .002). In the presence of lumbar endoleak, a 2.0-mm median increase (interquartile range, 0-5.0) in Dmax was seen (P ⫽ .045), and three of these eight patients (38%) underwent aneurysm sac growth of 5 mm or more. Secondary intervention. Changes in Dmax and the outcome of secondary intervention in patients with lumbar endoleak are shown in Table II. Only three of six patients (50%) had successful treatment with coil embolization compared with 14 of 19 patients (74%) before surgery (P ⫽ .344). Postoperative embolization took longer than preoperative embolization, typically in excess of 2 hours. Embolization was initially unsuccessful in one patient with a high-grade stenosis at the origin of the ileolumbar trunk.
108 Parry et al
Fig 1. Carbon dioxide digital subtraction angiograms. A, Lateral flush aortogram during embolization. Coils (black arrow) are clearly seen in IMA. B, Aortogram in anteroposterior projection after EVAR. Image shows retrograde filling of IMA (white arrows) from superior mesenteric artery via left colic artery. Embolization coils (black arrows) prevent retrograde perfusion of aneurysm sac.
The endoleak leak was eventually occluded via the contralateral ileolumbar trunk at a second embolotherapy session. In the patients in whom coil embolization was successful, reductions were seen in Dmax of ⫺1 mm, ⫺3 mm, and ⫺7 mm (Fig 4). Complications. No complications were seen related to preoperative embolization. A patient who underwent embolization with cyanoacrylate had a foot drop develop as the result of sciatic nerve palsy. DISCUSSION Reduction in size of the aneurysm sac is the only absolute indication of stent graft treatment success.10,14-16 In the presence of type II lumbar endoleak, we observed a small but statistically significant increase in sac diameter. In the recent
JOURNAL OF VASCULAR SURGERY July 2002
Fig 2. A, Image obtained during embolization of endoleak. Catheter (5F; black arrow) has been positioned in left hypogastric artery. Microcatheter (3F; white arrows) has been maneuvered via left iliolumbar trunk into right fourth lumbar artery. B, Carbon dioxide digital subtraction angiogram after embolization. Coils were deployed backwards from right fourth lumbar artery into left fourth lumbar artery (white arrows). Flow is preserved in left iliolumbar trunk and distal left fourth lumbar artery (black arrows).
series from Gould et al,17 four of 14 patients (29%) with type II endoleak had a more than 5 mm increase in sac diameter, which is similar to the 38% in our experience. Sac expansion is in keeping with the observation from Baum et al2 that type II endoleaks are perfused at near arterial pressure. In contrast, some studies have failed to show a significant change in aneurysm sac size in the presence of type II endoleak.10,14 Detection of small changes in aneurysm size can be problematic. Previous work has highlighted potential problems with repeatability of measurements of aneurysm diameter.18,19 Lederle et al19 have suggested that agreement to within 1 mm can be achieved when a consistent measuring technique is adopted. We have endeavoured to make measurements as objective and reproducible as possible by limiting reporting of CT scans to vascular radiologists and using mechanical callipers for all measurements. For prevention of bias, all measurements were made without knowledge of previous results. An audit of our technique has shown it to be accurate and reproducible.
JOURNAL OF VASCULAR SURGERY Volume 36, Number 1
Parry et al 109
Fig 3. Box whiskerplot shows median changes in sac size (Dmax) in patients with and without lumbar endoleak.
Type II endoleaks are frequently observed and account for between 18% and 30% of all endoleaks.20-22 In our series, despite the use of preoperative embolization, eight of 36 patients (22%) had lumbar endoleak develop. Without embolization, the observed endoleak rate likely would have been even greater. Increased rates of endoleak detection in this series might be the result of the combination of rigorous follow-up with dual phase CT and ultrasound scans, combined with a high index of suspicion. Our endoleak rate is comparable with the rate of recent work that has also used dual phase CT scan.11,17,23 Reduction of the incidence rate of type II endoleaks seems desirable. Recent work from Fan et al11 and Gorich et al12 has shown the influence of the number of patent branch vessels in the prediction of type II endoleak. In our series, no endoleaks were seen in the absence of patent side branches. This contrasts starkly with an endoleak rate of 62% in those with patent lumbar arteries. We successfully performed preoperative coil embolization in about three quarters of all patients without any procedural complications. Failure was usually attributable to unfavorable anatomic considerations, particularly the presence of luminal thrombus. In our series, adoption of a policy of preoperative embolization of all significant third and fourth lumbar arteries significantly reduced the incidence rate of lumbar endoleak. Most significantly sized lumbar vessels (19/ 21; 91%) were fourth lumbar. Common sense dictates that these larger vessels are more likely to remain patent after stent deployment. Gould et al17 reported no difference in the incidence rate of type II endoleak after preoperative embolization and concluded that the procedure was ineffective. Gould et al17 included small lumbar arteries in his “nonembolized” group. If these vessels were irrelevant, this would reduce the observed incidence rate of endoleak in the nonembolization group. Successful treatment in our series was defined as coil embolization of all target vessels. Gould et al17 included patients with incomplete embolization as a technical success. If significant lumbar vessels remain patent, than this will tend to increase the observed incidence rate of endoleak in the embolized group.
Fig 4. Late phase contrast-enhanced CT scans. A, Scan shows type II lumbar endoleak. Enhancement is seen within aneurysm sac (white arrows) both anterior and posterior to stent graft. Feeding right fourth lumbar artery (black arrows) can be seen adjacent to vertebral body. B, Corresponding image 3 months after coil embolization. Coils are seen in right fourth lumbar artery (black arrows), aneurysm has decreased in size, and perfusion of sac no longer exists.
All but one of the lumbar endoleaks we observed were delayed in onset, with detection after a median of 9 months of follow-up (range, 1 to 18 months). Some lumbar endoleaks may be present immediately after stent grafting but are only detected later with progressive enlargement. Equally, some endoleaks might be intermittent or occur with recanalization of an occluded artery. Late endoleaks likely represent a true de novo development, as observed by Gould et al,17 because none of the new endoleaks was demonstrable on review of the previous scans. That none of the type II endoleaks that we observed resolved spontaneously is surprising.24 In our experience, all of the lumbar endoleaks had demonstrable outflow vessels, suggesting that branch-to-branch circuits are necessary to maintain type II endoleaks. Our results suggest that these circuits have arisen in those patients in whom embolization was not attempted or was unsuccessful. Lumbar arteries arising from the aneurysm neck are likely to thrombose spontaneously because no space exists between the graft and the aortic wall.
110 Parry et al
Ermis et al25 have shown that successful embolization of patients with perigraft and collateral endoleaks can precipitate sac shrinkage. Once a type II endoleak has developed it can be treated with a variety of techniques.2,24-28 Embolization of established type II endoleak is a technically demanding and time-consuming procedure, with some patients needing more than one session to achieve occlusion. Our technical success rate with post-EVAR embolization was only three of six (50%), which is comparable with other series.25 After successful treatment of the endoleak, sac shrinkage occurred in every case, indicating significant pressurization of the sac. Alternative therapeutic options include the use of liquid embolic agents, but these increase the risk of the procedure. The sole neurologic complication of embolotherapy in our experience has been the patient with sciatic nerve palsy after glue embolization. In this patient, selective catheterization of the lumbar artery to deploy coils was impossible, and nontarget embolization occurred. Both sciatic and femoral nerve damage have been reported as a result of using particulate or liquid embolic agents in the pelvic circulation, particularly in the posterior division of the hypogastric artery.29 CONCLUSION Patent aortic side branches are associated with type II endoleaks and aneurysm growth. In our experience, preoperative coil embolization of IMA and lumbar artery side branches is technically possible in most patients and has prevented endoleak at a median follow-up of 2 years. Post-EVAR embolization of side branch endoleaks is technically more challenging with a lower success rate but can result in aneurysm shrinkage if successful. We thank Mr Patrick J. Kent, MCh, FRCSI, for his support and the provision of patients for this study. REFERENCES 1. White GH, Yu W, May J. Endoleak as a complication of endoluminal grafting of abdominal aortic aneurysms: classification, incidence, diagnosis and management. J Endovasc Surg 1997;4:152-68. 2. Baum RA, Carpenter JP, Cope C, Golden MA, Velazquez OC, Neschis DG, et al. Aneurysm sac pressure measurements after endovascular repair of abdominal aortic aneurysms. J Vasc Surg 2001;33:32-41. 3. Malina M, Ivancev K, Chuter TAM, Lindh M, Lanne T, Lindblad B, et al. Changing aneurysm morphology after endovascular grafting: relation to leakage or persistent perfusion. J Endovasc Surg 1997;4:23-30. 4. Schurink GWH, Aarts NJM, Wilde J, van Balen JM, Chuter TAM, Schultze Kool LJ, et al. Endoleakage after stent-graft treatment of abdominal aneurysm: implications on pressure and imaging—an in vitro study. J Vasc Surg 1998;28:234-41. 5. Gill K, Brennan J, Cheong CK. Factors influencing intrasac pressure following endovascular aortic aneurysm repair. Br J Surg 1998;85:1593. 6. Politz JK, Newman VS, Stewart MT. Late abdominal aortic aneurysm rupture after AneuRx repair: a report of three cases. J Vasc Surg 2000;31: 599-606. 7. Schurink GWH, Aarts NJM, van Balen JM, Chuter TAM, Schultze Kool LJ, van Bockel JH. Late endoleak after endovascular therapy for abdominal aortic aneurysm. Eur J Vasc Endovasc Surg 1999;17:448-50. 8. White RA, Donayre C, Walot I, Stewart M. Abdominal aortic aneurysm rupture following endoluminal graft deployment: report of a predictable event. J Endovasc Ther 2000;7:257-62.
JOURNAL OF VASCULAR SURGERY July 2002
9. Hinchcliffe RJ, Singh-Ranger R, Davidson IR, Hopkinson BR. Rupture of an abdominal aortic aneurysm secondary to type II endoleak. Eur J Vasc Endovasc Surg 2001;22:663-5. 10. Resch T, Ivancev K, Lindh M, Nyman U, Brunkwall J, Malina M, et al. Persistent collateral perfusion of abdominal aortic aneurysm does not lead to progressive change in aortic diameter. J Vasc Surg 1998;28:242-9. 11. Fan CM, Rafferty EA, Geller SC, Kauffman JA, Brewster DC, Cambria RP, et al. Endovascular stent graft in abdominal aortic aneurysms: the relationship between patent vessels that arise from the aneurysm sac and early endoleak. Radiology 2001;218:176-82. 12. Gorich J, Rilinger N, Sokiranski R, Soldner J, Kaiser W, Kramer S, et al. Endoleaks after endovascular repair of aortic aneurysms: are they predictable? Initial results. Radiology 2001;218:477-80. 13. Parry DJ, Waterworth A, Kessel D, Robertson I, Berridge DC, Scott DJA. Endovascular repair of an inflammatory abdominal aortic aneurysm complicated by aorto-duodenal fistulation with an unusual presentation. J Vasc Surg 2001;33:874-9. 14. Matsumura JS, Moore WS, for the Endovascular Technologies Investigators. Clinical consequences of peri-prosthetic leak after endovascular repair of abdominal aortic aneurysm. J Vasc Surg 1998;27:606-13. 15. Resch T, Ivancev K, Brunkwall J, Nirhov N, Malina M, Lindblad B. Midterm changes in aortic aneurysm morphology after endovascular repair. J Endovasc Ther 2000;7:279-85. 16. Wolf YG, Hill BB, Rubin GD, Fogarty FJ, Zairns CK. Rate of change of aneurysm diameter after endovascular repair. J Vasc Surg 2000;32:108-15. 17. Gould DA, McWilliams R, Edwards RD, Martin J, White D, Joekes E, et al. Aortic side branch embolisation before endovascular repair: incidence of type II endoleak. J Vasc Interv Radiol 2001;12:337-41. 18. Akkersdijk GJM, Puylaert CM, Coerkamp EG, DeVries AC. Accuracy of ultrasonographic measurements of infrarenal abdominal aortic aneurysm. Br J Surg 1994;81:376. 19. Lederle FA, Wilson SE, Johnson GR, Reinke DB, Littooy FN, Acher CW, et al. Variability in measurement of abdominal aortic aneurysms. J Vasc Surg 1995;21:945-52. 20. Schurink GWH, Aarts NJM, van Bockel JH. Endoleak after stent-graft treatment of abdominal aortic aneurysm: a meta-analysis of clinical studies. Br J Surg 1999;86:581-7. 21. Cupyers Ph, Buth J, Harris PL, Gevers E, Lahey R, on behalf of the EUROSTAR Collaborators. Realistic expectations for patients with stentgraft treatment of abdominal aortic aneurysms. Results of a European Multicentre Registry. Eur J Vasc Endovasc Surg 1999;17:507-16. 22. Vascular Surgical Society of Great Britain and Ireland. Fifth report on the Registry for Endovascular Treatment of Aneurysms (RETA). London: the Society; 2002. Available from http://www.vssgbi.org/ registries/5thretareport.pdf. 23. Golzarian J, Dussaussois L, Abada HT, Gevenois PA, van Gansbeke D, Ferreira J, et al. Helical CT of aorta after endoluminal stent-graft therapy: value of biphasic acquisition. Am J Roentgenol 1998;171:329-31. 24. Kato N, Semba CP, Dake MD. Embolisation of perigraft leaks after endovascular stent-graft treatment of aortic aneurysms. J Vasc Interv Radiol 1996;7:805-11. 25. Ermis C, Kramer S, Tomczac R, Palmer R, Kolokythas O, Schutz A, et al. Does successful embolisation of endoleaks lead to aneurysm sac shrinkage? J Endovasc Ther 2000;7:441-5. 26. Amesur NB, Zajko AB, Orons PD, Makaroun MS. Embolotherapy of persistent endoleaks after endovascular repair of abdominal aortic aneurysm with the Ancure-Endovascular Technologies Endograft System. J Vasc Interv Radiol 1999;10:1175-82. 27. LaBerge JM, Sawhney R, Wall SD, Chuter TA, Canto CJ, Wilson MW, et al. Retrograde catheterisation of the inferior mesenteric artery to treat endoleaks: anatomical and technical considerations. J Vasc Interv Radiol 2000;11:55-9. 28. Khilnani NM, Sos TA, Trost DW, Winchester PA, Jagust MB, Mitchell RS, et al. Embolisation of backbleeding lumbar arteries filling an aortic aneurysm sac after endovascular stent-graft placement. J Vasc Interv Radiol 1996;7:813-7. 29. Barton PP, Waneck RE, Karnel FJ, Ritschi P, Kramer J, Lechner GL. Embolisation of bone metasases. J Vasc Interv Radiol 1996;7:81-8. Submitted Sep 25, 2001; accepted Jan 17, 2002.
Endoleak is a complication specific to endovascular abdominal aortic aneurysm repair (EVAR) and has been defined as blood flow outside the stent graft but within the aneurysm sac.1 Aortic branch vessels (inferior mesenteric artery [IMA] and lumbar arteries) form potential communications between the aneurysm sac and the systemic circulation. Type II branch vessel endoleaks can transmit pressure into the aneurysm sac.2 Malina et al3 noted that even minor endoleak or collateral perfusion inhibits sac shrinkage. More recently, both in vitro and clinical studies have shown pulsatile systemic pressure within the aneurysm sac in the presence of type II endoleak.4,5 Cases of type II endoleak causing both sac growth and rupture also have been reported.6-9 Type II endoleaks have been considered relatively benign,10 but these results indicate their clinical importance. Recent evidence indicates the correlation between side branch patency and type II endoleak.11,12 Preoperative occlusion of these branches may prevent the subsequent development of type II endoleak. From the Departments of Vascular Surgerya and Radiology,b St James’s University Teaching Hospital, Leeds Teaching Hospitals. Competition of interest: nil. Reprint requests: Dr David Kessel, Department of Radiology, St James’s University Hospital, Beckett St, Leeds, LS9 7TF, United Kingdom (email: [email protected]). Copyright © 2002 by The Society for Vascular Surgery and The American Association for Vascular Surgery. 0741-5214/2002/$35.00 ⫹ 0 24/1/125023 doi:10.1067/mva.2002.125023
MATERIALS AND METHODS Design. This study was conducted as a prospective intervention study set in a university teaching hospital. Patient selection. The study population comprised consecutive patients with abdominal aortic aneurysm who underwent EVAR between September 1996 and April 2001. For determination of suitability, patients for EVAR underwent screening with dynamically enhanced spiral computed tomographic (CT) scan and digital subtraction angiography with a calibrated catheter. Patients suitable for EVAR had stent grafts sized according to manufacturer recommendations. All scans were reviewed by at least two vascular radiologists (DK, IR, JP), and patent aortic side branches (IMA and lumbar arteries) were noted. Patency was defined as evidence of perfusion of the vessel from the aortic lumen. For the purpose of the study, we defined a significantly sized lumbar vessel as being amenable to catheterization with a 4F or 5F catheter (ie, measuring approximately ⬎1.5 mm in diameter). Embolization was only attempted in lumbar arteries considered amenable to catheterization (ie, those measuring approximately ⬎1.5 mm in diameter). Preoperative coil embolization of aortic side branches. Preoperative embolization was performed before EVAR using 2-mm to 5-mm coils to occlude the target artery. In the IMA, coils were deployed proximal to the left colic artery (Fig 1). Currently, we attempt embolization in all patients with significantly sized patent lumbar arteries 105
JOURNAL OF VASCULAR SURGERY July 2002
106 Parry et al
lying outside the aneurysm neck. In our early experience, this was not routine, hence lumbar artery embolization was not attempted in all patients. Technical success was defined as angiographic occlusion of the IMA or all target lumbar arteries. Endovascular aortic aneurysm repair. EVAR was performed with general anesthesia in an interventional radiology suite with operating room conditions. Vanguard I (Boston Scientific, St Albans, UK), AneuRx (Medtronic, Watford, UK), and Excluder (Gore, Livingston, UK) stent grafts were used. Graft deployment was performed by a combined team of vascular radiologists and surgeons. Completion angiography was performed to check for endoleaks. Thirty-day morbidity and mortality were documented in all patients. Follow-up assessment. Patients were followed with plain radiographs and ultrasound and dual-phase contrast enhanced CT scans. Dynamically enhanced scans were commenced 30 seconds after the start of contrast infusion, and delayed scans taken 30 seconds later during venous phase enhancement. Scans were reviewed for evidence of patent branch arteries, endoleak, and maximum external diameter of the aneurysm sac. Diameters were measured by hand with mechanical calipers on the hard copy images. The greatest transverse diameter of the aneurysm sac (Dmax) was recorded. When the aneurysm sac was elliptic, the short axis measurement also was recorded. When an apparent increase was found in Dmax, comparison was made between scans to ensure that a comparable level was assessed and a reproducible measurement had been taken. Time of occurrence of type II endoleak was defined as the first instance of demonstrable sac perfusion on CT or ultrasound scan in the absence of a graft-related endoleak. Scans were performed before discharge and at 1, 3, 6, 12, 18, and 24 months and annually thereafter. Results were recorded on a database (DK, LD, DJAS). The site of origin of all leaks was confirmed with angiography, including, if necessary, selective ipsilateral hypogastric artery angiography. Endoleaks were classified as immediate (present from stent graft deployment), early (detected within 30 days of EVAR), or delayed (detected after 1 month during followup). Persistent endoleaks were defined as being detectable on at least two consecutive follow-up scans, at least 3 months apart. Secondary intervention. Embolization was attempted in patients with persistent lumbar endoleaks without shrinkage of the aneurysm sac. The ipsilateral hypogastric artery was catheterized with a 5F catheter, and access to the target lumbar artery was obtained with a coaxial microcatheter (Fig 2). Coils were deployed as close to the aneurysm sac as possible. In most cases, multiple Gianturco coils were deployed in the inflow vessel. If catheterization of the target vessel was impossible, then the feeding vessels were embolized. When collateral vessels perfused the adjacent lumbar arteries, the entire ileolumbar trunk was occluded. Glue (N-butyl 2-cyanoacrylate) embolization was performed in a single patient in whom catheterization of the right fourth lumbar artery, which fed the endoleak, was
Table I. Patent aortic side braches at presentation Patent vessels at presentation IMA All lumbar arteries* IMA and lumbar IMA and/or lumbar
No. 16 (45%) 21 (53%)* 10 (25%) 27 (67.5%)
*Of 21 patients with patent lumbar arteries, 19 (90%) had a patent fourth lumbar artery; 7 (33%) had a patent fourth and second/third lumbar artery; and 2 (10%) had a patent third lumbar artery alone.
impossible. Particulate embolization was not used. Success was defined as abolition of the endoleak on angiography and subsequent imaging studies. Outcome measures. The primary outcome measure was the incidence of type II endoleak. Secondary outcome measures were the technical success rate of preoperative and postoperative embolization and change in the Dmax. Statistical analysis. All data were analyzed with SPSS for Windows version 9.0 (SPSS, Inc, Chicago, Ill). Categoric data were analyzed with a two-tailed Fisher exact test for the difference in proportions of type II endoleak in those cases with successful embolization as compared with those with patent side branches at baseline. Nonparametric data were analyzed with Wilcoxon matched pairs signed rank test for changes in Dmax. RESULTS Forty patients (34 male, six female) underwent treatment with EVAR, with Vanguard I, AneuRx, and Excluder stent grafts deployed in five, 33, and two patients, respectively. Median age was 75.6 years (range, 54 to 92 years). Twenty-seven patients (67.5%) had patent IMA or lumbar arteries (Table I shows the prevalence of patent branch vessels on preoperative imaging). Preoperative embolization. Nineteen of the 27 patients with patent aortic side branches underwent attempted preoperative embolization. The procedures typically took between 30 and 60 minutes and were without complication. Technical success was defined as angiographic occlusion of the IMA or all target lumbar arteries. All successful lumbar vessel embolization involved coil deployment in the larger fourth and third lumbar arteries. The median number of lumbar vessels embolized was two (range, 1 to 5). Embolization was successful in 14 patients (74%), and 21 target vessels (78%) were occluded. Endovascular aneurysm repair. Successful aneurysm exclusion was achieved in 38 of 40 patients (95%). One immediate type III endoleak with no demonstrable outflow vessel was seen. The leak was the result of a tiny “pinholelike” defect in the fabric of the right iliac limb. A single type II endoleak was associated with a solitary patent lumbar artery. Both settled spontaneously by 30 days and did not recur during follow-up. The 30-day mortality rate was four of 40 (10%). Two deaths occurred after resections for colonic malignant disease 2 weeks after EVAR. The need for major abdominal surgery had been an indication for
JOURNAL OF VASCULAR SURGERY Volume 36, Number 1
Parry et al 107
Table II. Patients with lumbar endoleak: origin, changes in sac size, and outcome of secondary intervention
Case 1 2 3 4 5 6 7 8 Median
Source inflow 3 outflow
Time when detected (mo)
⌬ Dmax (mm)
Secondary embolization
Further interventions or events
L L4 3 R L4 L L3 3 R L4 L L4 3 R L4 R L4 3 L L4 R L4 3 L L3/4 L L4 3 R L4 R L4 3 L L4. L L4 3 R L4
18
⫹6
Successful embolization‡
6
0
12
⫹5
6
⫹2
12
0
1
⫹2
12
⫹1
3
⫹5
Failed embolization Awaiting embolization Died before intervention† Failed embolization* Successful embolization Failed embolization Successful embolization Successful embolization
9
⫹2
⌬ Dmax2 –3
0 –7 Died during follow-up† –1 Died during follow-up†
Success rate 3/6 (50%)
*Unable to traverse aneurysm sac. Inflow vessel only embolized. † All deaths were unrelated to stent grafting. ‡ Laproscopic clipping failed, followed by successful second attempt at embolization. L3, third lumbar artery; L4, fourth lumbar artery; R, right; L, left; ⌬ Dmax, change in maximum sac diameter before secondary intervention; ⌬ Dmax2, change in sac diameter after secondary intervention.
EVAR in these patients, and in each case, recovery after the aneurysm repair was uneventful. The third patient died of cholecystitis associated with cardiac failure on the 30th postoperative day. The fourth died of small bowel infarction 48 hours after EVAR. A subsequent postmortem examination revealed extensive atheroma in the suprarenal aorta, and the superior mesenteric artery was patent. Patchy small intestinal infarction was thought to have arisen from massive embolization of atheromatous debris during stent deployment. These latter two patients were octogenarians and were classified as unfit for open aneurysm repair. In addition, both patients had large aneurysms with short markedly angulated proximal aneurysm necks. Both procedures were complex and necessitated placement of additional proximal cuffs. These four patients were all excluded from the subsequent analysis. Follow-up. The median follow-up period for all patients (n ⫽ 36) was 24 months (range, 6 to 48 months), and follow-up periods of the patients with and without embolization were 24 months (range, 6 to 48 months) and 30 months (range, 18 to 36 months), respectively. The incidence rate of lumbar endoleak was eight of 36 (22%) during follow-up (Table II). The median time of lumbar endoleak detection was 9 months (range, 1 to 18 months). All endoleaks persisted on at least two consecutive scans. Seven endoleaks were detected with CT scan, and one with duplex scan. All endoleaks arose from the third or fourth lumbar arteries and had a patent outflow source. Two patients (6%) had a type I endoleak develop (one proximal, one distal), and one patient (3%) a type III endoleak. One patient (3%) with an inflammatory aneurysm had an aor-
toduodenal fistula develop at 6 months that needed stent graft replacement with an in situ silver impregnated graft.13 These patients were all excluded from statistical analysis of changes in sac size. Effect of preoperative embolization on incidence of type II endoleak. No type II endoleaks were seen among patients who had spontaneously occluded aortic side branches at baseline (n ⫽ 13) or who underwent successful IMA or lumbar embolization (n ⫽ 14). No IMA endoleaks were found. Eight of 13 patients (62%) with patent lumbar arteries had a subsequent lumbar endoleak develop, compared with none of the eight patients (0%) who had undergone preoperative lumbar embolization (P ⫽ .006, Fisher exact test). Changes in maximum sac diameter at follow-up. Changes in Dmax are illustrated in Fig 3. In the absence of lumbar endoleak, a mean reduction of 3.0 mm (interquartile range, 10.5-0) was seen in Dmax (P ⫽ .002). In the presence of lumbar endoleak, a 2.0-mm median increase (interquartile range, 0-5.0) in Dmax was seen (P ⫽ .045), and three of these eight patients (38%) underwent aneurysm sac growth of 5 mm or more. Secondary intervention. Changes in Dmax and the outcome of secondary intervention in patients with lumbar endoleak are shown in Table II. Only three of six patients (50%) had successful treatment with coil embolization compared with 14 of 19 patients (74%) before surgery (P ⫽ .344). Postoperative embolization took longer than preoperative embolization, typically in excess of 2 hours. Embolization was initially unsuccessful in one patient with a high-grade stenosis at the origin of the ileolumbar trunk.
108 Parry et al
Fig 1. Carbon dioxide digital subtraction angiograms. A, Lateral flush aortogram during embolization. Coils (black arrow) are clearly seen in IMA. B, Aortogram in anteroposterior projection after EVAR. Image shows retrograde filling of IMA (white arrows) from superior mesenteric artery via left colic artery. Embolization coils (black arrows) prevent retrograde perfusion of aneurysm sac.
The endoleak leak was eventually occluded via the contralateral ileolumbar trunk at a second embolotherapy session. In the patients in whom coil embolization was successful, reductions were seen in Dmax of ⫺1 mm, ⫺3 mm, and ⫺7 mm (Fig 4). Complications. No complications were seen related to preoperative embolization. A patient who underwent embolization with cyanoacrylate had a foot drop develop as the result of sciatic nerve palsy. DISCUSSION Reduction in size of the aneurysm sac is the only absolute indication of stent graft treatment success.10,14-16 In the presence of type II lumbar endoleak, we observed a small but statistically significant increase in sac diameter. In the recent
JOURNAL OF VASCULAR SURGERY July 2002
Fig 2. A, Image obtained during embolization of endoleak. Catheter (5F; black arrow) has been positioned in left hypogastric artery. Microcatheter (3F; white arrows) has been maneuvered via left iliolumbar trunk into right fourth lumbar artery. B, Carbon dioxide digital subtraction angiogram after embolization. Coils were deployed backwards from right fourth lumbar artery into left fourth lumbar artery (white arrows). Flow is preserved in left iliolumbar trunk and distal left fourth lumbar artery (black arrows).
series from Gould et al,17 four of 14 patients (29%) with type II endoleak had a more than 5 mm increase in sac diameter, which is similar to the 38% in our experience. Sac expansion is in keeping with the observation from Baum et al2 that type II endoleaks are perfused at near arterial pressure. In contrast, some studies have failed to show a significant change in aneurysm sac size in the presence of type II endoleak.10,14 Detection of small changes in aneurysm size can be problematic. Previous work has highlighted potential problems with repeatability of measurements of aneurysm diameter.18,19 Lederle et al19 have suggested that agreement to within 1 mm can be achieved when a consistent measuring technique is adopted. We have endeavoured to make measurements as objective and reproducible as possible by limiting reporting of CT scans to vascular radiologists and using mechanical callipers for all measurements. For prevention of bias, all measurements were made without knowledge of previous results. An audit of our technique has shown it to be accurate and reproducible.
JOURNAL OF VASCULAR SURGERY Volume 36, Number 1
Parry et al 109
Fig 3. Box whiskerplot shows median changes in sac size (Dmax) in patients with and without lumbar endoleak.
Type II endoleaks are frequently observed and account for between 18% and 30% of all endoleaks.20-22 In our series, despite the use of preoperative embolization, eight of 36 patients (22%) had lumbar endoleak develop. Without embolization, the observed endoleak rate likely would have been even greater. Increased rates of endoleak detection in this series might be the result of the combination of rigorous follow-up with dual phase CT and ultrasound scans, combined with a high index of suspicion. Our endoleak rate is comparable with the rate of recent work that has also used dual phase CT scan.11,17,23 Reduction of the incidence rate of type II endoleaks seems desirable. Recent work from Fan et al11 and Gorich et al12 has shown the influence of the number of patent branch vessels in the prediction of type II endoleak. In our series, no endoleaks were seen in the absence of patent side branches. This contrasts starkly with an endoleak rate of 62% in those with patent lumbar arteries. We successfully performed preoperative coil embolization in about three quarters of all patients without any procedural complications. Failure was usually attributable to unfavorable anatomic considerations, particularly the presence of luminal thrombus. In our series, adoption of a policy of preoperative embolization of all significant third and fourth lumbar arteries significantly reduced the incidence rate of lumbar endoleak. Most significantly sized lumbar vessels (19/ 21; 91%) were fourth lumbar. Common sense dictates that these larger vessels are more likely to remain patent after stent deployment. Gould et al17 reported no difference in the incidence rate of type II endoleak after preoperative embolization and concluded that the procedure was ineffective. Gould et al17 included small lumbar arteries in his “nonembolized” group. If these vessels were irrelevant, this would reduce the observed incidence rate of endoleak in the nonembolization group. Successful treatment in our series was defined as coil embolization of all target vessels. Gould et al17 included patients with incomplete embolization as a technical success. If significant lumbar vessels remain patent, than this will tend to increase the observed incidence rate of endoleak in the embolized group.
Fig 4. Late phase contrast-enhanced CT scans. A, Scan shows type II lumbar endoleak. Enhancement is seen within aneurysm sac (white arrows) both anterior and posterior to stent graft. Feeding right fourth lumbar artery (black arrows) can be seen adjacent to vertebral body. B, Corresponding image 3 months after coil embolization. Coils are seen in right fourth lumbar artery (black arrows), aneurysm has decreased in size, and perfusion of sac no longer exists.
All but one of the lumbar endoleaks we observed were delayed in onset, with detection after a median of 9 months of follow-up (range, 1 to 18 months). Some lumbar endoleaks may be present immediately after stent grafting but are only detected later with progressive enlargement. Equally, some endoleaks might be intermittent or occur with recanalization of an occluded artery. Late endoleaks likely represent a true de novo development, as observed by Gould et al,17 because none of the new endoleaks was demonstrable on review of the previous scans. That none of the type II endoleaks that we observed resolved spontaneously is surprising.24 In our experience, all of the lumbar endoleaks had demonstrable outflow vessels, suggesting that branch-to-branch circuits are necessary to maintain type II endoleaks. Our results suggest that these circuits have arisen in those patients in whom embolization was not attempted or was unsuccessful. Lumbar arteries arising from the aneurysm neck are likely to thrombose spontaneously because no space exists between the graft and the aortic wall.
110 Parry et al
Ermis et al25 have shown that successful embolization of patients with perigraft and collateral endoleaks can precipitate sac shrinkage. Once a type II endoleak has developed it can be treated with a variety of techniques.2,24-28 Embolization of established type II endoleak is a technically demanding and time-consuming procedure, with some patients needing more than one session to achieve occlusion. Our technical success rate with post-EVAR embolization was only three of six (50%), which is comparable with other series.25 After successful treatment of the endoleak, sac shrinkage occurred in every case, indicating significant pressurization of the sac. Alternative therapeutic options include the use of liquid embolic agents, but these increase the risk of the procedure. The sole neurologic complication of embolotherapy in our experience has been the patient with sciatic nerve palsy after glue embolization. In this patient, selective catheterization of the lumbar artery to deploy coils was impossible, and nontarget embolization occurred. Both sciatic and femoral nerve damage have been reported as a result of using particulate or liquid embolic agents in the pelvic circulation, particularly in the posterior division of the hypogastric artery.29 CONCLUSION Patent aortic side branches are associated with type II endoleaks and aneurysm growth. In our experience, preoperative coil embolization of IMA and lumbar artery side branches is technically possible in most patients and has prevented endoleak at a median follow-up of 2 years. Post-EVAR embolization of side branch endoleaks is technically more challenging with a lower success rate but can result in aneurysm shrinkage if successful. We thank Mr Patrick J. Kent, MCh, FRCSI, for his support and the provision of patients for this study. REFERENCES 1. White GH, Yu W, May J. Endoleak as a complication of endoluminal grafting of abdominal aortic aneurysms: classification, incidence, diagnosis and management. J Endovasc Surg 1997;4:152-68. 2. Baum RA, Carpenter JP, Cope C, Golden MA, Velazquez OC, Neschis DG, et al. Aneurysm sac pressure measurements after endovascular repair of abdominal aortic aneurysms. J Vasc Surg 2001;33:32-41. 3. Malina M, Ivancev K, Chuter TAM, Lindh M, Lanne T, Lindblad B, et al. Changing aneurysm morphology after endovascular grafting: relation to leakage or persistent perfusion. J Endovasc Surg 1997;4:23-30. 4. Schurink GWH, Aarts NJM, Wilde J, van Balen JM, Chuter TAM, Schultze Kool LJ, et al. Endoleakage after stent-graft treatment of abdominal aneurysm: implications on pressure and imaging—an in vitro study. J Vasc Surg 1998;28:234-41. 5. Gill K, Brennan J, Cheong CK. Factors influencing intrasac pressure following endovascular aortic aneurysm repair. Br J Surg 1998;85:1593. 6. Politz JK, Newman VS, Stewart MT. Late abdominal aortic aneurysm rupture after AneuRx repair: a report of three cases. J Vasc Surg 2000;31: 599-606. 7. Schurink GWH, Aarts NJM, van Balen JM, Chuter TAM, Schultze Kool LJ, van Bockel JH. Late endoleak after endovascular therapy for abdominal aortic aneurysm. Eur J Vasc Endovasc Surg 1999;17:448-50. 8. White RA, Donayre C, Walot I, Stewart M. Abdominal aortic aneurysm rupture following endoluminal graft deployment: report of a predictable event. J Endovasc Ther 2000;7:257-62.
JOURNAL OF VASCULAR SURGERY July 2002
9. Hinchcliffe RJ, Singh-Ranger R, Davidson IR, Hopkinson BR. Rupture of an abdominal aortic aneurysm secondary to type II endoleak. Eur J Vasc Endovasc Surg 2001;22:663-5. 10. Resch T, Ivancev K, Lindh M, Nyman U, Brunkwall J, Malina M, et al. Persistent collateral perfusion of abdominal aortic aneurysm does not lead to progressive change in aortic diameter. J Vasc Surg 1998;28:242-9. 11. Fan CM, Rafferty EA, Geller SC, Kauffman JA, Brewster DC, Cambria RP, et al. Endovascular stent graft in abdominal aortic aneurysms: the relationship between patent vessels that arise from the aneurysm sac and early endoleak. Radiology 2001;218:176-82. 12. Gorich J, Rilinger N, Sokiranski R, Soldner J, Kaiser W, Kramer S, et al. Endoleaks after endovascular repair of aortic aneurysms: are they predictable? Initial results. Radiology 2001;218:477-80. 13. Parry DJ, Waterworth A, Kessel D, Robertson I, Berridge DC, Scott DJA. Endovascular repair of an inflammatory abdominal aortic aneurysm complicated by aorto-duodenal fistulation with an unusual presentation. J Vasc Surg 2001;33:874-9. 14. Matsumura JS, Moore WS, for the Endovascular Technologies Investigators. Clinical consequences of peri-prosthetic leak after endovascular repair of abdominal aortic aneurysm. J Vasc Surg 1998;27:606-13. 15. Resch T, Ivancev K, Brunkwall J, Nirhov N, Malina M, Lindblad B. Midterm changes in aortic aneurysm morphology after endovascular repair. J Endovasc Ther 2000;7:279-85. 16. Wolf YG, Hill BB, Rubin GD, Fogarty FJ, Zairns CK. Rate of change of aneurysm diameter after endovascular repair. J Vasc Surg 2000;32:108-15. 17. Gould DA, McWilliams R, Edwards RD, Martin J, White D, Joekes E, et al. Aortic side branch embolisation before endovascular repair: incidence of type II endoleak. J Vasc Interv Radiol 2001;12:337-41. 18. Akkersdijk GJM, Puylaert CM, Coerkamp EG, DeVries AC. Accuracy of ultrasonographic measurements of infrarenal abdominal aortic aneurysm. Br J Surg 1994;81:376. 19. Lederle FA, Wilson SE, Johnson GR, Reinke DB, Littooy FN, Acher CW, et al. Variability in measurement of abdominal aortic aneurysms. J Vasc Surg 1995;21:945-52. 20. Schurink GWH, Aarts NJM, van Bockel JH. Endoleak after stent-graft treatment of abdominal aortic aneurysm: a meta-analysis of clinical studies. Br J Surg 1999;86:581-7. 21. Cupyers Ph, Buth J, Harris PL, Gevers E, Lahey R, on behalf of the EUROSTAR Collaborators. Realistic expectations for patients with stentgraft treatment of abdominal aortic aneurysms. Results of a European Multicentre Registry. Eur J Vasc Endovasc Surg 1999;17:507-16. 22. Vascular Surgical Society of Great Britain and Ireland. Fifth report on the Registry for Endovascular Treatment of Aneurysms (RETA). London: the Society; 2002. Available from http://www.vssgbi.org/ registries/5thretareport.pdf. 23. Golzarian J, Dussaussois L, Abada HT, Gevenois PA, van Gansbeke D, Ferreira J, et al. Helical CT of aorta after endoluminal stent-graft therapy: value of biphasic acquisition. Am J Roentgenol 1998;171:329-31. 24. Kato N, Semba CP, Dake MD. Embolisation of perigraft leaks after endovascular stent-graft treatment of aortic aneurysms. J Vasc Interv Radiol 1996;7:805-11. 25. Ermis C, Kramer S, Tomczac R, Palmer R, Kolokythas O, Schutz A, et al. Does successful embolisation of endoleaks lead to aneurysm sac shrinkage? J Endovasc Ther 2000;7:441-5. 26. Amesur NB, Zajko AB, Orons PD, Makaroun MS. Embolotherapy of persistent endoleaks after endovascular repair of abdominal aortic aneurysm with the Ancure-Endovascular Technologies Endograft System. J Vasc Interv Radiol 1999;10:1175-82. 27. LaBerge JM, Sawhney R, Wall SD, Chuter TA, Canto CJ, Wilson MW, et al. Retrograde catheterisation of the inferior mesenteric artery to treat endoleaks: anatomical and technical considerations. J Vasc Interv Radiol 2000;11:55-9. 28. Khilnani NM, Sos TA, Trost DW, Winchester PA, Jagust MB, Mitchell RS, et al. Embolisation of backbleeding lumbar arteries filling an aortic aneurysm sac after endovascular stent-graft placement. J Vasc Interv Radiol 1996;7:813-7. 29. Barton PP, Waneck RE, Karnel FJ, Ritschi P, Kramer J, Lechner GL. Embolisation of bone metasases. J Vasc Interv Radiol 1996;7:81-8. Submitted Sep 25, 2001; accepted Jan 17, 2002.