Author: Mustafa Halawa, MD, Benjamin Patterson, MD, Peter Holt, MD, Matt Thompson, MD, Ian M. Loftus, MD Author Affiliations: From St. George’s Vascular Institute, London, England. Correspondence: Ian Loftus, MD, St. George’s Vascular Institute, St. George’s Hospital, Blackshaw Road, London, England SW170QT Manuscript submitted September 12, 2008, provisional acceptance given October 15, 2008, and accepted November 6, 2008. Disclosure: The authors report no financial relationships or conflicts of interest regarding the content herein. _______________________________ Abstract Thoracic aortic pathology confers specific challenges to stent graft design and performance. The indications for endovascular intervention are expanding, and the complexity of cases is increasing. Graft design has evolved, and several new iterations of endografts have been introduced into clinical practice. While most devices are probably suitable for straightforward cases, extra care should be exercised in device selection for complex pathology. The assessment of these new grafts is difficult, due to the changing indications for their use and a lack of large-scale clinical trials. Although clinical trials provide the highest level of evidence, they often utilize strict anatomical criteria for inclusion and may not be representative of clinical practice. Despite their inherent shortcomings, multicenter registry data offer a good opportunity for evaluating new endovascular grafts and techniques, as large numbers of cases can be collected in relatively short periods of time. Introduction Endovascular repair of thoracic aortic pathology has gained support due to evidence from registries demonstrating acceptable operative mortality and paraplegia rates.1,2 The United States Food and Drug Administration (FDA) approved the commercial use of thoracic endografts in the treatment of thoracic aortic aneurysms (TAA) in 2005.3 Since then, several trials have shown endovascular repair to be a safe and effective alternative to open surgery, with clear outcome benefits.4–6 In comparison with earlier devices, the latest generation of stent grafts have improved the ease of use and reliability, but the question of the long-term durability of endografts remains unanswered. This may be addressed by the analysis of registries that have been created at the request of advisory bodies such as the National Institute for Health and Clinical Excellence in the United Kingdom. Organizations such as these have had particular interest in this technology, as the cost of development has been significant and the current market cost of the devices remains a major consideration in a healthcare system with limited resources. The demands of vascular specialists have outstripped the provision of endograft design and manufacture, as the indications for stent grafting have expanded and the complexity of cases undertaken has increased.7–10 This has exposed some of the technical limitations of the early endograft designs. The thoracic aorta poses specific difficulties in endograft delivery, deployment, and fixation.11 Procedures such as hybrid endovascular repair of the thoraco-abdominal aorta with visceral and supra-aortic trunk debranching and repair of the disease of the aortic arch typify these challenges (Figure 1). It is difficult to determine if the choice of device significantly impacts a patient’s outcome, due to a lack of studies that examine graft choice. This could be explained, to a degree, by the fact that a combination of patient factors, different aortic morphologies, and anatomical factors may lend themselves to a particular device. What Particular Problems Does Thoracic Endografting Pose? During deployment, a long length of devices must often negotiate heavily diseased arteries. This can cause trauma to the aortic wall and increases the risk of embolization of atheromatous debris into the brain or distally to the limbs. Once the device has been manipulated into place, endografts can fail to conform to the curved shaped of the aortic arch, due to their relatively inflexible configuration (Figure 2). If this happens, grafts can be subject to migration, collapse, and pseudo-coarctation. If a device is not fixed securely, long-term durability may be jeopardized, and although data are not comprehensive, some recent studies have suggested there may be an incidence of late aortic-related deaths following thoracic endografting.18 Methods such as hooks and radial force have been used to facilitate fixation of the devices, and this produces a certain amount of stress within the wall of the aorta. Damage to the intima-provoking aortic dissection and erosion through the aortic wall are possible consequences. These factors should be considered when using certain devices for specific indications, although, as yet, there is no definite evidence to support this practice. Another difficulty faced when placing a graft in the thoracic aorta is the accuracy of deployment, particularly in the aortic arch. Torque placed on the delivery system by the operator outside the patient may not be translated to device movement in the proximal aorta. This may result in forceful manipulation of the device before deployment. If the device is not deployed accurately, proximal endoleaks or occlusion of proximal aortic branches may occur. Selection of Endograft There are a number of devices currently available with variable amounts of clinical data to support their use. Some devices have relatively unique features that differentiate them from others. Gore TAG Endoprosthesis Device description. The Gore TAG endoprosthesis (W.L. Gore, Flagstaff, Arizona) is an ePTFE graft with an exoskeleton provided by a nitinol stent (Figure 3). Radial force fixation is provided by a spring flare, and additional security is provided by a sealing cuff at each end. Radio-opaque, gold-marker bands at the base of the flares aid with positioning. There was no longitudinal-supporting strut; it was removed when the first generation Gore stent graft was modified. It is delivered using the Gore Introducer and deployed using Gore’s SIM-PULLTM release system. The TAG endograft is available in diameters of 26–40 mm, enabling the treatment of patients with landing zone diameters of 23–37 mm in the descending thoracic aorta. The graft is supplied in 3 lengths: 10, 15, and 20 cm. Several overlapping grafts may be deployed to treat longer lesions. An ePTFE sleeve constrains the device with a plastic obturator on the leading edge of the delivery system. Delivery initiates from the central portion of the device, extending in both directions simultaneously. Although this facilitates rapid deployment, some feel that this can reduce the accuracy of final placement, especially in the aortic arch. A tri-lobed balloon has been designed to secure fixation after device deployment with a design that allows continuous blood flow through it during inflation. Pros and cons. The TAG endograft is extremely easy to deploy. Its rapid deployment has obvious advantages in situations such as aortic transection, when rapid exclusion of the aortic tear is paramount. However, “the all or nothing” nature of the delivery system is also its major disadvantage. For this reason, the accuracy of deployment is sometimes suboptimal. When the landing zones are close to the supra-aortic branches or the celiac artery, the device is probably not ideal. Furthermore, the stent-graft does not conform well around acutely angulated aortic arches, such as in young patients. This results in the graft sitting up in the arch, with the potential for the collapse of the proximal end. One other important drawback of the system is the requirement of a relatively large introducer sheath. This fact should be kept in consideration when the iliac arteries are of borderline caliber. Data from trials and registries. In 2005, the FDA approved the routine use of the Gore TAG device in the treatment of TAA following the results of its phase II pivotal trial.12 This was a prospective multi-center study that took place between 1999 and 2001 and compared the outcome in 94 patients who underwent open repair to a group of 142 patients who were treated with the Gore TAG device. In 98% of cases, the graft was implanted successfully, with 32% of patients suffering from major complications, including a 4% stroke rate and a 3% paraplegia rate. At 30 days postoperatively, there was a 2.1% death rate for endovascular repair, compared to 11.7% for those in the open surgery group. However, at 2-year follow up, the all-cause mortality did not differ significantly (24% and 26% for endovascular and open surgery, respectively). The incidence of major adverse events (MACE) was significantly lower in patients from the endovascular group, compared with those who underwent open surgical repair (42% versus 77%, respectively). After 3 years, 5 patients required endovascular re-intervention, and 1 patient had an open surgical revision. Although 24 stent fractures were detected, no patients died due to device-related complications. In 2003, the confirmatory trial was performed after the removal of the fracture-prone longitudinal wires to compare the safety and early results of the modified device with the original device.13 This was a prospective single-arm, non-randomized trial performed on 51 patients at 11 test centers. The MACE rates were compared with the control group from the original trial (94 patients who underwent open repair). At 30 days, 12% of patients in the endovascular group experienced MACE, although no deaths or major device-related events were observed. In 2008, the 5-year follow-up results of the Gore TAG trials were published, reporting an aneurysm-related mortality for TAG patients at 2.8%, compared with 11.7% for open controls (P = 0 008).4 No differences in all-cause mortality were noted, with 68% of TAG patients and 67% of open controls surviving up to 5 years (P = 0 43). MACE at 5 years were significantly reduced in the TAG group; 57.9% versus 78.7% (P = 0 001). Endoleaks in the TAG group decreased from 8.1% at 1 month to 4.3% at 5 years. Five TAG patients underwent major aneurysm-related re-interventions at 5 years (3.6%), including 1 arch aneurysm repair for type 1 endoleak and migration, 1 open conversion, and 5 endovascular procedures for endoleaks in 3 patients. There were fewer secondary procedures not directly related to aneurysm repair in the TAG versus the open repair group at 5 years (15.0% vs 31.9%, P = 0 01). For TAG patients, sac size at 60 months decreased in 50% and increased in 19%, compared with the 1-month baseline. Comparison with the modified low-porosity device at 24 months showed sac increase in 12.9% of original versus 2.9% in modified grafts. At 5 years, there were no ruptures, 1 migration, no collapse, and 20 instances of fracture in 19 patients, all before the revision of the TAG graft. The TAG system has been used to treat several different types of thoracic aortic pathology. This could explain why higher complication rates have been reported in association with this practice, and caution has been advised when using this particular device in complex disease of the aortic arch.14,15 Medtronic Talent System Device description. The Talent System consists of a self-expanding nitinol stent with an inlaid woven polyester graft (Medtronic, Santa Rosa, California) (Figure 4).16 It incorporates a bar throughout its length for longitudinal strength while maintaining the ability to negotiate tortuous and diseased vessels. Proximally, a “closed web” configuration or an open “FreeFlo” bare stent variety is available. The bare-stent grafts are designed for situations where deployment across the origin of the left subclavian or common carotid artery are unavoidable. This helps to ensure persistent antegrade flow postoperatively. The bare-wire configuration can also be incorporated into the distal stent graft, allowing flow into the visceral branches of the abdominal aorta should these be covered during exclusion of disease areas. Delivery systems are available in profiles sized 22–25 Fr, and the actual device is available in diameters of 22–46 mm. Data from trials and registries. The Talent Thoracic Retrospective Registry was a multi-center collaboration that collected data from patients treated with endovascular devices in Europe between 1996 and 2004.17 Of the 457 patients that underwent repair with the Talent Thoracic Stent Graft, 137 (29.9%) had atherosclerotic aneurysms, 180 (39.4%) had dissections, and 85 (18%) had traumatic lesions. There were 113 emergency and 344 elective cases. Technical success was achieved in 97.8%, with the failure in the remaining 2.2% attributed mainly to inadequate peripheral access vessels or to difficulties deploying the device. Mortality rates were 7.9% and 4.1% for emergency and elective cases, respectively. Stroke and paraplegia rates were 3.7% and 1.7%, respectively. Mean follow up was 2 years, and the late mortality rate was 8.5%. A third of these deaths were aortic related. The VALOR study was a multi-center, non-randomized trial that compared the results of endovascular treatment with retrospective open surgical data.6 The endovascular group consisted of 195 patients with TAA. All but one graft was deployed successfully, and 30-day follow up revealed a 2.1% risk of operative mortality and a 41% chance of MACE. There was less than a 1.5% chance of paraplegia and a 3.6% chance of stroke. At 12 months, there was a 16.1% aneurysm-related mortality rate and a 12.2% endoleak rate. The endovascular group had a statistically better outcome in all acute outcomes, except for vascular outcomes associated with the thoracic aorta. The Talent stent graft was therefore deemed safe for use in patients who were candidates for open surgery. Medtronic Valiant System Device description. The Valiant endovascular graft (Figure 5) (Medtronic, Santa Rosa, California) is a third generation endoprosthesis that further develops the Talent device, featuring improved design of the stent, delivery sheath, graft configuration, deployment methodology, and markers.11 The Xcelerant delivery system has an integrated handle similar to the Talent AAA stent graft, providing a mechanical advantage that requires less force to be used during deployment. Improved design of the sheath system has resulted in reduced stretching and better trackability. The proximal stent has a “free flow” and “closed web” configuration, and is available in tapered or straight designs. The distal fixation component may also be closed or have a bare spring, and increased flexibility has been achieved by removing the connecting bar while preserving columnar support through narrower stent spacing. A wider range of dimensions is available, with lengths up to 227 mm and diameters from 24–46 mm. These longer grafts may have the added benefit of reducing modular junctional endoleaks and have the potential to reduce the number of times delivery systems have to be manipulated through diseased vessels. The proximal stent includes an 8-peak spring to provide proximal fixation, allowing closer apposition to the aortic side walls and distribution of the radial force over more apices (Figure 6). Table 1 compares the different features of the Medtronic Talent with the Valiant devices. Pros and cons. The Xcelerant delivery system of the Valiant device enables easy, controlled, and accurate deployment, which is particularly useful in the aortic arch. The device conforms very well to the aortic arch and, at present, is probably the best device available for patients with marked angulation at the junction of the arch and descending aorta. It is also very flexible and can advance through challenging, diseased, iliac arteries. The Valiant endograft has the widest range of lengths and diameters among its competitors. However, a range of smaller sizes for younger patients and the production of a narrower delivery system would be beneficial. Data from trials and registries. A recently published registry from 7 European centers examined outcomes in 180 patients who underwent endovascular repair with the Valiant system.18 Preemptive open surgical procedures prior to covering aortic branches included three ascending aortic bypasses to the brachiocephalic and left common carotid arteries, 17 carotid-carotid, 14 carotid-subclavian bypasses, 18 visceral bypasses, and 11 access procedures. More than one device was deployed in 50% of procedures, compared to 30% in the Talent study, despite the longer length of the Valiant device. This is an indicator that the Valiant group was comprised of patients with more challenging anatomy. Despite this, the device was deployed successfully in 95% of cases. Failure was ascribed to difficult access in three cases and deployment problems in the remaining five. There were 13 immediate device-related complications that included one iliac perforation, 7 Type I endoleaks, 3 Type III endoleaks, and 2 retrograde Type A dissections. There were no immediate open conversions. The overall 30-day mortality was 7.2%, with elective and emergency mortality of 6.8% and 8.6%, respectively. Stroke and paraplegia rates were both 3.8%. However, this included those who underwent hybrid endovascular repair of thoraco-abdominal aneurysms with visceral debranching, and the mortality and paraplegia rates were 28% and 18%, respectively for this subgroup. If these high-risk procedures were excluded, the mortality and paraplegia rates were 3% and 1.3%, respectively. Based on these results, it seems that the Valiant stent graft represents a significant design improvement over previous designs. The success of this device has encouraged attempts to treat increasingly complex aortic pathology. However, more information regarding long-term durability needs to be collected to define the precise indications in these complex cases. Cook Zenith TX2 System Device description. The Zenith TX2 TAA (Cook Inc., Bloomington, Indiana) is a two-part stent graft system, consisting of a Dacron graft covered by stainless steel and modified Gianturco Z-stents. Both the proximal and distal components are tubular stent grafts and have enhanced fixation through multiple 5 mm barbs: the proximal covered stent of the proximal component features caudally oriented barbs protruding through the fabric (Figure 7), while the distal uncovered bare stent of the distal component has cranially oriented barbs, which fix the graft at the distal end and prevent proximal migration (Figure 8). A recent modification to facilitate the treatment of thoracic aortic dissection is a bare-metal aortic-stent component. Deployment is achieved by withdrawal of an external sheath. The proximal part of the system is held in place by a 3-wire system. The distal end of the graft is held in place by a constraining cap to prevent unintentional release of the device. The full length of the graft material is supported throughout by supportive stents along the length of the fabric component of the device. The Zenith TX2 components are available in diameters, ranging from 22–42 mm and lengths from 10.8–21.6 cm. The Z-Trak delivery systems for the Zenith TX2 device have profiles between 20 and 22 Fr. It should be noted that these refer to the inner diameter of the sheath. Pros and cons. The presence of trigger wires that prevent initial full expansion of the endograft provides additional control when accuracy of deployment is paramount. Therefore, the TX2 device is well suited for patients when the landing zones are close to the supra-aortic branches and the celiac artery. The delivery system is the lowest profile of the 3 main thoracic endografts, which provides a potential advantage for young patients with aortic transaction, or when the access vessels are narrow. On the other hand, the device does not conform to curves as well as some of the other endografts. Also, the proximal fixation barbs would theoretically make the TX2 less suitable for patients with aortic dissection because of the potential for trauma to the fragile intima. Data from trials and registries. In 2005, interim results of the first 100 consecutive repairs in a larger planned series using the first generation TX1 and TX2 systems were reported.19 Most of these cases consisted of atherosclerotic, degenerative aortic aneurysms. After one year of follow up, the all-cause mortality rate was 17%, with the majority of these deaths being aneurysm-related (14%). Stroke and paraplegia rates of 3% and 6%, respectively, were reported. Within this period, the endoleak rate was 6%, with secondary interventions required in 15%, and stent migration demonstrated in 6%. A prospective, non-randomized, multinational trial involving 42 centers and 230 patients undergoing repair of descending TAA with the Zenith TX2 device was recently reported.5 Graft deployment was successful in all but two patients. Of these, 59.5% required a two-piece device and 40.5% required only one. The 30-day survival rate was comparable to open repair (98.1% and 94.3%, respectively), as was the re-intervention rate, but the morbidity score at 30 days was significantly lower in the endovascular group. Measures of clinical utility suggested that endovascular repair was superior. Follow up at 12 months revealed growth of the aneurysm in 7.1% of patients, endoleak in 3.9%, and migration in 2.8%. This trial has confirmed that the TX2 device is a safe alternative to open repair. Table 2 compares the different features of the three main available stent grafts. Summary Thoracic aortic pathology creates specific demands on stent graft design and performance, and as indications for endovascular intervention are expanding, the complexity of cases is increasing further. This is likely to add impetus to new developments. In addition to degenerative aneurysms, traumatic injury, dissection, and ulcers of the aorta can be treated using endovascular strategies. Over time, the design of stent grafts has evolved, and new generations of thoracic endografts have been introduced into clinical practice. While most devices are probably suitable for straightforward cases, device selection for complex pathology must be undertaken with due care. Assessment of strengths and weaknesses of these new grafts is difficult, due to the different indications for their use and a lack of large-scale clinical trials. Trials organized by bodies such as the FDA often stipulate strict anatomical criteria for inclusion, and this does not necessarily reflect standard clinical practice. Data from multicenter registries may eventually provide the best solution to this problem, since large amounts of data can be collected relatively simply. Comparative trials are difficult to design because of the relative pros and cons of each for different pathologies and the wide range of pathologies treated in modern endovascular practice. That said, it should be possible to perform direct comparisons of stent graft performance for individual pathologies both in the short and long term.