Endograft Technology: Highlights of the Past 10 Years

II-192 J ENDOVASC THER REVIEW Endograft Technology: Highlights of the Past 10 Years Thomas J. Fogarty, MD; Frank R. Arko, MD; and Christopher K. Zarins, MD Division of Vascular Surgery, Stanford University Medical Center, Stanford, California, USA. The past decade has seen the evolution of an exciting technology that has changed forever the treatment of aortic aneurysmal disease. From rather crude homemade stent-grafts constructed in the surgical suite to elegant commercially manufactured devices in a variety of configurations and sizes, the aortic endograft has experienced a meteoric rise in popularity to become a beneficial, minimally invasive therapy that can obviate the risk of rupture and death. There are now 3 approved endovascular devices on the market for infrarenal abdominal aortic aneurysm repair, and it is likely that additional and improved devices will become available in the future. This review revisits the developmental history of the aortic endograft, noting the ongoing refinements that have arisen from our experiences with the growing population of stent-graft patients. Although research continues to search for solutions to the problems of endoleak and migration, long-term results even with the earlier second and third-generation devices are better than has been achieved with open surgical repair. J Endovasc Ther Key words: abdominal aortic aneurysm, stent-graft, device design, stent, graft material, endoleak, rupture The treatment of aortic aneurysms has changed dramatically over the past 10 years as a result of innovative endovascular strategies and the development and application of new technology. Juan Parodi s 1 demonstration that blood flow could be excluded from an abdominal aortic aneurysm (AAA) using a transluminally placed surgical graft aroused worldwide interest in the potential of endovascular aneurysm repair. The benefits to the patient of a minimally invasive transfemoral approach that did not require aortic crossclamping were obvious. However, the technological challenges of designing a suitable endoluminal device that could exclude the aneurysm from the circulation and withstand the formidable biomechanical forces of pulsatile aortic blood flow were felt by some to be insurmountable compared to the security of sutured surgical grafts. Some viewed the new experimental endovascular approach as doomed to failure, considering the longstanding experience with open surgery and the comfort level that vascular surgeons had with the procedure. Nonetheless, the past decade has witnessed remarkably rapid technological advances in endovascular device design and delivery systems as a result of collaborations among vascular surgeons, engineers, interventional radiologists, and industry. A number of devices have been evaluated in multicenter controlled clinical trials, and all have shown that endovascular repair is effective in preventing aneurysm rupture with less periprocedural morbidity, faster patient recovery, Address for correspondence and reprints: Christopher K. Zarins, MD, Division of Vascular Surgery, 300 Pasteur Drive, Room H-3642, Stanford University Medical Center, Stanford, CA 94305-5450 USA. Fax: 1-650-498-6044; E-mail: zarins@stanford.edu 2004 by the INTERNATIONAL SOCIETY OF ENDOVASCULAR SPECIALISTS Available at www.jevt.org

J ENDOVASC THER II-193 and earlier return to function than open surgical repair. In 1999, only 8 years after Parodi s first successful endovascular AAA procedure, the Food and Drug Administration (FDA) granted market approval for two endovascular devices, the AneuRx 2 (Medtronic Vascular, Santa Rosa, CA, USA) and the Ancure 3 (Guidant Corporation, Indianapolis, IN, USA). Although the Ancure endograft is no longer on the market, the AneuRx device has gained wide clinical acceptance, with more than 50,000 patients treated worldwide. Favorable long-term clinical results now extend to more than 8 years. 4 More recently, two additional devices have gained FDA approval, the Excluder 5 (W.L. Gore & Associates, Flagstaff, AZ, USA) and the Zenith 6 (Cook Inc., Bloomington, IN, USA). Thus, in the short span of 10 years, stentgraft repair has revolutionized the treatment of aneurysms and become the procedure of choice for suitable patients with infrarenal AAA. While endovascular repair of thoracic aortic aneurysms (TAA) has been performed for more than 10 years with good clinical success, 7 technological challenges have delayed FDA approval and widespread clinical application in the US. It is anticipated that endovascular devices for thoracic and thoracoabdominal aneurysms will become available in the near future. DEVELOPMENT OF ENDOVASCULAR DEVICES The principle of endovascular repair is similar to open surgical repair, in which exclusion of the aneurysm from the aortic bloodstream is achieved by attaching a prosthetic graft to the nonaneurysmal aorta proximally and distally. Early endovascular devices were constructed by suturing polyester prosthetic grafts to selfexpanding or balloon-expandable stents that served to anchor the graft to the nonaneurysmal aorta by applying radial force to the aortic neck. These early homemade devices were designed to treat selected AAAs and descending TAAs, either as tube grafts or as aortomonoiliac exclusions with contralateral iliac occlusion and femorofemoral bypass. These first-generation devices had large delivery systems and significant limitations. The need for better endograft designs with smaller diameter delivery systems and the ability to deal with the complex anatomy of aortoiliac aneurysms quickly became apparent. Early design considerations included bifurcated unibody grafts versus modular systems, differing graft materials and weave densities, self-expanding versus balloon-expandable stents, exoskeleton versus endoskeleton, radial force versus hook fixation, columnar support versus flexibility, suprarenal versus infrarenal fixation, delivery system diameter, flexibility and design, and percutaneous versus open delivery. A variety of endovascular devices were developed by multiple manufacturers utilizing different strategies of construction, fixation, and deployment. The early Endovascular Technologies (EVT) aortic tube graft, first deployed by Wesley Moore in 1994, 8 had limited usefulness since most aortic aneurysms involved the aortic bifurcation. 9 This device was replaced by a unibody bifurcated endograft design that required a complex delivery system to deploy the contralateral iliac limb. Simpler delivery systems were possible using a modular component design and bifemoral device deployment. In addition, the EVT graft had no radial support in the iliac limbs and thus was prone to extrinsic graft compression or kinking in diseased and tortuous iliac arteries. Stent and hook fractures, as well as fabric density and durability, were important issues in early devices. Stent-graft construction options included placing the stent inside (endoskeleton) or outside (exoskeleton) of the fabric graft, and a variety of attachment mechanisms were used to join the stent to the fabric. In areas of severe angulation, an endoskeletal configuration could result in abrasion of the polyester fabric against the ends of the metallic stents, with perforation of the graft material. Such fabric tears could lead to device failure and aneurysm rupture. 10 The exoskeletal structure mitigated this problem and also increased surface friction between the stent and the aortic wall, allowing tissue in-growth into the interstices of the stent to assist in anchoring the device. Suprarenal fixation mechanisms, with and without hooks, were successfully incor-

II-194 J ENDOVASC THER Figure 1Evolution and improvement of the AneuRx stent-graft. (A) The early prototype stiff bifurcation module (left) was replaced with a segmented stent structure (right) that added flexibility to the device. (B) Fabric weave density was increased (C) in order to reduce porosity and enhance durability. (D) The blunt-nosed delivery system that was used during the clinical trial was replaced with an improved tapered tip delivery system (E) in 2002. porated into some devices to improve fixation in angulated aortic necks. 6,11 A discussion of the design variables in stent-graft construction and the search for the ideal endograft design were summarized by Robert Allen several years ago. 12 Using these concepts, a modular stent-graft design with a simple, intuitive delivery system was designed by Fogarty Engineering more than 10 years ago. This modular endograft design was a new conceptual approach at that time, utilizing both radial fixation force and longitudinal columnar support to maintain position in the aorta. Mechanical and bench testing and animal implantations were begun in 1994, 13 and the first human implantation was performed in 1996. A controlled phase II multicenter clinical trial was completed in 1998, resulting in FDA approval of the AneuRx stent-graft in 1999. This stent-graft design remains one of the most widely used endografts for AAA repair worldwide, with durable long-term results. 14 Early clinical experiences and endograft device failures resulted in rapid advances and evolution in device design. For example, it was learned that the early prototype AneuRx bifurcation module was too stiff and did not conform well in angulated aortoiliac aneurysms. Thus, in 1997, the bifurcation module was modified to a segmented stent design that allowed flexibility of the device (Fig. 1A). In addition, a more tightly woven fabric with reduced porosity (Fig. 1 B,C) was introduced in 1998 in order to eliminate transgraft flow, which was commonly seen on completion angiograms. Furthermore, it was observed that retraction of the runners in the delivery system after deployment caused excessive friction between the delivery system sheath and the bifurcated endograft module, which could displace the stent-graft and necessitate implantation of aortic extender cuffs. The delivery system was subsequently improved to significantly reduce the friction in the runner system, thus preventing intraoperative device migration. This new delivery system, which was introduced in 2002, included a tapered nosecone to replace the original blunt bullet configuration (Fig. 1 D,E), which also lowered the outer diameter of the introducer sheath to 21-F. The current AneuRx stent-graft is now in its 4 th generation of device design. Improve-

J ENDOVASC THER II-195 ments have addressed issues of deployment, delivery, density of fabric weave, and ease of use, while maintaining the same basic qualities of the ideal endovascular device. Future changes of this device will include improved radiopaque markers, a longer main body, an increased variety of sizes, and an integrated deployment handle. Similar technological advances and improvements have occurred with other endovascular devices in the past decade. For example, the polytetrafluoroethylene (PTFE) fabric of the Excluder endoprosthesis has recently been modified to reduce its porosity in an attempt to decrease or eliminate the problem of aneurysm enlargement associated with this graft material. The Cook Zenith is also modifying its outer stent design by increasing the space between the stents in an attempt to increase the flexibility of this device. ENDOVASCULAR SKILLS AND TRAINING AND IMPROVED IMAGING Equally important to the rapid technological advances in device design was the need to provide the vascular surgical community with endovascular skills and training. Endovascular repair was a new conceptual strategy for treating aortic aneurysms quite unlike open surgical repair. While many vascular surgeons, particularly those involved with the early clinical trials, had some endovascular experience in peripheral occlusive disease, this was an entirely new approach to aneurysms for most vascular surgeons. It required an intimate knowledge of pre and postoperative imaging modalities, including cross-sectional computed tomography (CT) with 3-dimensional reconstruction. It also required the skills to work under fluoroscopy with indirect imaging, to use guidewires and catheters, and to perform angiographic imaging and interpretation. It required conceptualizing nonsurgical aneurysm exclusion strategies that were quite different from the well developed techniques of open aortic aneurysm repair. Thus, it was no surprise that early efforts were met with resistance and skepticism from some segments of the traditional vascular surgery community. When the first prospective clinical trials of endovascular aneurysm repair began in 1995 (EVT-EGS system) and 1996 (AneuRx), few vascular surgeons had the experience and endovascular skills needed to deploy endovascular devices. For most clinical sites in these multicenter trials, this was the first institutional experience with endovascular aneurysm repair. Thus, a steep learning curve existed in both the trials, which required assembling endovascular teams, imaging facilities, equipment, and support systems. Collaborations between vascular surgeons and interventional radiologists and/or cardiologists facilitated the successful deployment of the devices, which required femoral artery exposure and repair. Early imaging systems often were suboptimal. For example, we began endovascular aneurysm repair in the operating room at Stanford in 1996 using a portable 9-inch OEC 9400 C-arm fluoroscope and a table top ruler to help with positioning of the device. However, despite these early limitations, the results of endovascular aneurysm repair were remarkably good, with a 98% to 99% deployment success rate and a significant reduction in operative time and blood loss, with shorter hospital stays, fewer complications, and faster recovery compared to open surgical repair. 2,3,5,6 After a decade of experience and exposure, most vascular surgeons have acquired the skills needed to successfully treat patients with endovascular grafts. The unequivocal benefit of endovascular repair over open surgery has recently been confirmed in two European prospective randomized trials. Both the British and Dutch studies showed that endovascular aneurysm repair had a 3 to 4-fold reduction in operative mortality compared to open surgical repair. 15,16 Since completion of the first endograft clinical trials in 1999, there have been marked improvements in imaging systems, techniques, and endovascular skills, which allow greater precision in endograft deployment and positioning. Magnification images, C-arm angulation and rotation, moveable fluoroscopy tables, and effective catheters, guidewires, and sheaths are now used routinely by vascular surgeons who have acquired impressive catheters skills. The education of the vascular surgery community was facilitated by manufacturer-supported training programs in endovascular

II-196 J ENDOVASC THER aneurysm repair, which was a condition of FDA approval of the endovascular devices in 1999. These educational efforts have been very effective in improving the skills of vascular surgeons and have favorably impacted the outcomes using all devices. THE LEARNING CURVE: OUTCOME ANALYSIS DRIVEN In addition to the technical aspects of patient selection and device deployment, much has been learned during the past 10 years about long-term device performance. Much of the learning curve with respect to long-term outcomes has occurred as a result of ongoing clinical and image surveillance of patients who had been enrolled into the Ancure and AneuRx clinical trials. During the course of these studies, the primary endpoint of success was believed to be complete aneurysm exclusion with elimination of endoleak (flow in the aneurysm sac). It was assumed that if there was no endoleak, the risk of aneurysm rupture had been eliminated. To this day, some consider endoleak to be the primary measure of long-term success of endovascular repair. However, after the clinical trials were completed, reports of aneurysm rupture appeared, prompting much discussion and a detailed failure analysis of patients who experienced rupture. This analysis, published in May 2000, 17 showed that the 7 patients who experienced post-implant rupture among almost 1200 patients treated during the AneuRx trial all had evidence of poor stent-graft fixation proximally, distally, or at the junction gate. Furthermore, evidence for such poor fixation could be seen on the immediate postimplant CT scans and plain abdominal radiographs in all the rupture patients. Moreover, the defects were visible on follow-up imaging studies for up to 2 years before aneurysm rupture occurred, which suggested that rupture might have been prevented if poor fixation had been (1) avoided by more precise device deployment and use of extender cuffs at the initial procedure or (2) corrected by performing a secondary procedure to secure fixation with extender cuffs when imaging studies showed evidence of poor fixation (Fig. 2). This information on the importance of fixation was incorporated into all training courses and was disseminated to physicians beginning in 2000, thus reducing the risk of rupture. It is also interesting to note that the risk of rupture was higher among patients treated with the early stiff bifurcated module that was used at the beginning of the clinical trial. 18 The lack of flexibility of this prototype device increased the likelihood of poor fixation in areas of neck angulation and confirmed the importance of device fixation. Although the stiff device was used only in the very early phases of the clinical trial and has never been commercially available, these ruptures remain in the reported long-term outcomes with the AneuRx device. 14 Similar early learning curve ruptures have also been observed with the Ancure and other stent-graft devices, but they usually have not been included in long-term outcome reporting. 19 The longterm risk of rupture for patients treated with the current FDA-approved devices is very low, although occasional ruptures do still occur, and ongoing patient monitoring is needed. A surprise finding of the AneuRx rupture analysis was that 5 of the 7 rupture patients had no evidence of endoleak on repeated postprocedural CT studies prior to rupture. The 2 patients with endoleak had poor or absent fixation of the device to the iliac artery, which resulted in prominent type I endoleaks; however, both patients refused treatment. These observations raised questions regarding the significance of endoleaks and what it means if a patient has an endoleak. Is this really a useful predictive endpoint, as we had assumed during the clinical trial? In order to answer these questions, we utilized objective Core Laboratory documentation of endoleak on serial postprocedural CT scans from 383 patients in phase II of the multicenter trial. We found that the presence or absence of an endoleak was not a predictor of the endpoints of aneurysm rupture, surgical conversion, stent-graft migration, or patient survival. 20 The presence of an endoleak was related to an increased likelihood of aneurysm enlargement of 5 mm or more. However, most patients with enlargement have not required treatment, and sac expansion, with or without endoleak, was not related to an increased risk of rupture or decreased surviv-

J ENDOVASC THER II-197 Figure 2Successful endovascular repair yesterday (A) and today (B). (A) This patient was considered to have a successful endovascular repair since the aneurysm was completely excluded from the circulation, and there was no endoleak. However, the device is positioned far below the renal arteries and barely into the iliac arteries, thus predisposing the patient to stent-graft migration and possible rupture. (B) Current deployment techniques seek to position the device just below the renal arteries with extension of the limbs to the level of the hypogastric arteries, thus providing long-term stability. al. 21 These findings have now been confirmed by others, and type II endoleaks in and of themselves are no longer viewed as strong indicators of risk of rupture. LONG-TERM DURABILITY: STENT-GRAFT MIGRATION Long-term durability of endovascular aneurysm repair requires maintenance of endograft position and seal against the relentless downward force of pulsatile aortic blood flow. A variety of fixation mechanisms have been employed, including infrarenal fixation, suprarenal fixation, hooks, barbs, radial force, and longitudinal columnar support. Regardless of the fixation mechanism, migration has been noted with all current endovascular devices. The migration rate is 2% at 1 year using each of the FDA-approved devices. 5,22,23 In a longer term analysis of migration among AneuRx clinical trial patients, we found that the same factors that were of importance in preventing aneurysm rupture are also important in preventing stent-graft migration. During a 5-year period, migration was identified in 94 (8%) of 1119 patients. However, the migration rates varied widely among the 13 clinical sites (0% to 30%, Fig. 3), which suggests that factors other than the device played a major role in determining migration, since all sites used the same device. Indeed, multivariate analysis showed that low deployment of the device below the renal arteries was the most important predictor of subsequent migration. For each millimeter increase in the distance between the renal arteries and the top of the stent-graft on postdeployment CT scans, the risk of subsequent migration increased by 6%. 21 Thus, several sites, even some with large clinical experiences, consistently deployed the device low in the aortic neck, predisposing to migration. Further analysis revealed that iliac fixation length was also of great importance in preventing migration. Given the early phase of the learning

II-198 J ENDOVASC THER Figure 3Migration rates according to clinical site in the AneuRx clinical trial. The migration rate varied from 0% to 30% among the 13 clinical sites, suggesting that factors other than the device played a major role in determining migration. curve, the less refined imaging techniques, and the primary focus on endoleak as a procedural endpoint during the clinical trial, it is not surprising that the device was deployed well below the renal arteries and barely into the iliac arteries in some patients. If no endoleak was seen on completion angiography, the procedure was deemed to be a success. The fact that short initial proximal and distal fixation lengths predisposed to late stentgraft migration was appreciated only years later. Stent-graft migration is now considered to be preventable by positioning the device just below the renal arteries and extending the device to the level of the hypogastric artery bilaterally (Fig. 2B). Similar long-term analysis is needed for all endovascular devices. LONG-TERM OUTCOME AND PROSPECTS FOR THE FUTURE This past decade has witnessed a remarkable evolution and transformation in our approach to the treatment of aortic aneurysms. Even with the learning curve and early clinical trial results included, the long-term outcomes of endovascular infrarenal AAA repair are excellent. Five-year Kaplan-Meier results are now available for all 1193 patients treated during the AneuRx clinical trial from 1996 to 1999, including early prototype devices and the high-risk and off-protocol patients. 4 They show results that are considerably better than open surgical repair, with a remarkable 97% freedom from rupture and aneurysm-related death and a 91% freedom from surgical conversion at 5 years. Similar favorable longterm results using other FDA-approved devices may be expected in the future. In the next decade we can expect continuing improvements in endovascular device design to include the treatment of juxtarenal, suprarenal, and thoracoabdominal aneurysms. Clinical trials of devices to treat descending TAAs are nearing completion and should have FDA approval in the near future. The role of open surgical repair of aortic aneurysms will diminish but not disappear. As we gain experience and skill in endovascular aortic aneurysm repair, surgeons will be challenged to maintain their skills in open repair. REFERENCES 1. Parodi JC, Palmaz JC, Barone HD. Transfemoral intraluminal graft implantation for abdominal aortic aneurysms. Ann Vasc Surg. 1991;5: 491 499. 2. Zarins CK, White RA, Schwarten D, et al.

J ENDOVASC THER II-199 AneuRx stent graft versus open surgical repair of abdominal aortic aneurysms: multicenter prospective clinical trial. J Vasc Surg. 1999;29: 292 308. 3. Makaroun MS, Chaikof E, Naslund T, et al. Efficacy of a bifurcated endograft versus open repair of abdominal aortic aneurysms: a reappraisal. J Vasc Surg. 2002;35:203 210. 4. Zarins CK. The new gold standard? The AneuRx clinical trial at 8 years: lessons learned following the US AneuRx clinical trial from 1996 to 2004. Endovascular Today. 2004; July Supplement:7 11. 5. Matsumura JS, Brewster DC, Makaroun MS, et al. A multicenter controlled clinical trial of open versus endovascular treatment of abdominal aortic aneurysm. J Vasc Surg. 2003;37:262 271. 6. Greenberg RK, Lawrence-Brown M, Bhandari G, et al. An update of the Zenith endovascular graft for abdominal aortic aneurysms: initial implantation and mid-term follow-up data. J Vasc Surg. 2001;33:S157 164. 7. Dake MD, Miller DC, Semba CP, et al. Transluminal placement of endovascular stent-grafts for the treatment of descending thoracic aortic aneurysms. N Engl J Med. 1994;331:1729 1734. 8. Moore WS, Vescera CL. Repair of abdominal aortic aneurysm by transfemoral endovascular graft placement. Ann Surg. 1994;220:331 341. 9. Moore WS, Rutherford RB. Transfemoral endovascular repair of abdominal aortic aneurysm: results of the North American EVT phase 1 trial. J Vasc Surg. 1996;23:543 553. 10. Beebe HG, Cronenwett JL, Katzen BT, et al. Results of an aortic endograft trial: impact of device failure beyond 12 months. J Vasc Surg. 2001;33:S55 63. 11. Criado FJ, Fairman RM, Becker GJ. Talent LPS AAA stent graft: results of a pivotal clinical trial. J Vasc Surg. 2003;37:709 715. 12. Allen RC, White RA, Zarins CK, et al. What are the characteristics of the ideal endovascular graft for abdominal aortic aneurysm exclusion? J Endovasc Surg. 1997;4:195 202. 13. Beygui RE, Kinney EV, Pelc LR, et al. Prevention of spinal cord ischemia in an ovine model of abdominal aortic aneurysm treated with a selfexpanding stent-graft. J Endovasc Surg. 1999; 6:278 284. 14. Zarins CK; AneuRx Clinical Investigators. The US AneuRx Clinical Trial: 6-year clinical update 2002. J Vasc Surg. 2003;37:904 908. 15. Greenhalgh RM, Brown LC, Kwong GP, et al. Comparison of endovascular aneurysm repair with open repair in patients with abdominal aortic aneurysm (EVAR trial 1), 30-day operative mortality results: randomised controlled trial. Lancet. 2004;364:843 848. 16. Prinssen M, Verhoeven ELG, Buth J, et al. A randomized trial comparing conventional and endovascular repair of abdominal aortic aneurysms. N Engl J Med. 2004;351:1607 1618. 17. Zarins CK, White RA, Fogarty TJ. Aneurysm rupture after endovascular repair using the AneuRx stent graft. J Vasc Surg. 2000;31:960 970. 18. Zarins CK, White RA, Moll FL, et al. The AneuRx stent graft: four-year results and worldwide experience 2000. J Vasc Surg. 2001; 33:S135 145. 19. Bernhard VM, Mitchell RS, Matsumura JS, et al. Ruptured abdominal aortic aneurysm after endovascular repair. J Vasc Surg. 2002;35: 1155 1162. 20. Zarins CK, White RA, Hodgson KJ, et al. Endoleak as a predictor of outcome after endovascular aneurysm repair: AneuRx multicenter clinical trial. J Vasc Surg. 2000;32:90 107. 21. Zarins CK, Bloch DA, Crabtree T, et al. Aneurysm enlargement following endovascular aneurysm repair: AneuRx clinical trial. J Vasc Surg. 2004;39:109 117. 22. Zarins CK, Bloch DA, Crabtree T, et al. Stent graft migration after endovascular aneurysm repair: importance of proximal fixation. J Vasc Surg. 2003;38:1264 1272. 23. Sternbergh WC3rd, Money SR, Greenberg RK, et al. Influence of endograft oversizing on device migration, endoleak, aneurysm shrinkage, and aortic neck dilation: results from the Zenith Multicenter Trial. J Vasc Surg. 2004;39:20 26.