Chapter 54. Endovascular Therapy for the Treatment of Thoracic Aortic Disease

Patients with thoracic aortic disease are a difficult population to treat, because they frequently consist of an aged population with multiple comorbidities. The modern surgical treatment of thoracic aortic diseases began in the 1950s when successful treatment using segmental resection and graft replacement was reported by Swan, Lam, DeBakey, and Etheredge.1–3 Thereafter, DeBakey and Cooley reported the first successful repair of an ascending aortic aneurysm using cardiopulmonary bypass.4 Our understanding of the pathophysiology and natural history of thoracic aortic disease has evolved, which has expanded our treatment choices.5,6 In addition, improvements in diagnostic capabilities, surgical techniques, and perioperative care have resulted in improved outcomes, even as the risk profile has increased. Nonetheless, operative intervention in this patient population frequently results in substantial mortality and long-term morbidity.7,8 The concept of using endovascular techniques to treat patients with thoracic aortic disease emerged a decade ago, propelled by the desire to avoid surgical risk as well as to induce reconstructive modeling of the diseased aorta by initiating a natural healing process through exclusion and depressurization of the aneurysmal sac.9 In an effort to improve outcomes in the treatment of patients with thoracic aortic disease, endovascular stent-graft technology has rapidly followed applications on the abdominal aorta.10,11 Originally devised for high-risk patients with multiple comorbidities, thoracic stent-graft applications are being expanded to young and old patients with a variety of pathologies, including thoracic aortic aneurysms, aortic dissections, intramural hematomas, penetrating atherosclerotic ulcers, and thoracic aortic trauma.12–19 Initial reports using these endovascular stent-grafts have been encouraging, but long-term outcomes are unknown, and the necessity for long-term follow-up, with its attendant expense, has raised serious concern.20–22

Endovascular stent-graft technology was initially envisioned for use in abdominal aortic aneurysms.23 Introduced by Parodi, balloon-expandable stents attached to the ends of a vascular tube graft were used to exclude the aneurysm sac. There were several attractive features of this concept, including the introduction of the device from a peripheral site, eliminating the necessity for an invasive laparotomy, the avoidance of aortic cross-clamping and its requisite physiologic perturbations, and minimizing respiratory complications. Last, hospital stay and recovery time could be potentially shortened.

At Stanford University Medical Center, a collaborative effort between interventional radiologists and cardiovascular surgeons proved highly synergistic, and resulted in the manufacture and clinical use of thoracic stent grafts. Work had commenced years earlier with the use of uncovered stents for the repair of aortic dissections in an animal model. The stent grafts were manufactured using self-expanding Gianturco Z stents (Cook Co., Bloomington, IN), which were fastened together and then covered with a woven Dacron graft (Meadox-Boston Scientific, Natick, MA; Fig. 54-1). Institutional review board (IRB) approval was initially obtained for a high-risk study using endovascular stent-grafts for the treatment of thoracic aortic aneurysms in patients who were deemed not to be surgical candidates.24 A total of 13 patients underwent transluminal endovascular grafting of thoracic aortic aneurysms with a mean diameter of 6.1 cm. The stent grafts, custom-designed for each patient, were constructed of self-expanding stainless steel stents covered with woven Dacron grafts. Placement of these stents was successful in all patients with thrombosis of the aneurysm surrounding the stent occurring in 12 of the 13 patients. As reported, at 1 year there were no deaths, paraplegia, stroke, distal embolization, or infection.24 It was therefore concluded that these preliminary results demonstrated that endovascular stent-graft repair was safe in highly selected patients.

This feasibility trial led to the extension of the IRB approval for the treatment of 103 patients with thoracic aortic aneurysms.25 Of these 103 patients, 60% were unsuitable candidates for conventional open surgical repair and therefore deemed inoperable. Again, these patients underwent repair of a descending thoracic aortic aneurysm using "homemade" or first-generation stent grafts fabricated from self-expanding Z stents covered with woven Dacron tube graft. Complete aneurysm thrombosis was achieved in 83% of patients. Early mortality was 9% and was significantly associated with a history of cerebrovascular accidents and myocardial infarctions. Major perioperative morbidity included paraplegia in three patients, cerebrovascular accidents in seven patients, and respiratory insufficiency in 12 patients. Treatment failure occurred in 38 of the 103 patients, and five patients required late operative therapy for endoleaks associated with aneurysm enlargement. Actuarial survival was 81% at 1 year and 73% at 2 years. Given the high-risk nature of this patient population, these first-generation results were deemed satisfactory. It was, however, recognized that mortality and morbidity occurred frequently and that long-term follow-up was necessary to fully define the efficacy of an endovascular approach in thoracic aortic aneurysm therapy. Subsequently, in 2004, midterm results were reported for these 103 patients treated with the first-generation stent grafts.26 Overall actuarial survival was dismal; 82, 49, and 27% at 1, 5, and 8 years, respectively. However, the survival of the potentially operable candidates was 93 and 78% at 1 and 5 years, respectively, as compared with 74 and 31% at 1 and 5 years, respectively, in those patients designated as inoperable. In patients judged not to be surgical candidates, life expectancy, despite endovascular therapy, was therefore quite bleak, and has raised concerns whether any surgical therapy is appropriate for these patients. Further results revealed that 11 of the 103 patients suffered late aortic rupture at the site of endovascular treatment. This was a very sobering finding considering that open surgical graft replacement has been associated with durable long-term results with only a negligible late hazard of anastomotic problems. However, it must be remembered that this study involved the use of a relatively primitive first-generation device, with a fairly steep learning curve.

Extending the use of the endovascular stent grafts to the treatment of complicated acute aortic dissections of the descending thoracic aorta, Dake and associates reported their findings in the New England Journal of Medicine in 1999.27 Again these stents were the first-generation "homemade" devices described in the preceding. Placement of the stents across the primary entry tears was technically successful in all patients, with correction of malperfusion. Complete thrombosis of the false lumen occurred in 79%. The early mortality rate was 16%, which reflected late referral, with established intestinal gangrene, and another patient with Ehlers-Danlos syndrome, perhaps a poor candidate for endovascular repair. Favorable clinical results persisted out to a mean follow-up of 13 months.

These pioneering efforts at the Stanford University Medical Center established some interesting concepts. Namely, that these were complex patients with complex aortic problems, and that endograft therapy for aneurysmal disease was effective, potentially with reduced morbidity, but with uncertain long-term durability. Improved results could likely be obtained with more sophisticated devices, and endograft repair could effectively reverse malperfusion syndromes in complicated type B aortic dissections.

Thoracic Aortic Aneurysms

Approximately 50% of all thoracic aortic aneurysms are located in the descending aorta; these aneurysms commonly arise at the level of the left subclavian artery and are often atherosclerotic in nature.28 The size-rupture correlation has been demonstrated by studying the natural history of these aneurysms as reported by Clouse and associates, using the Olmstead County database, in which thoracic aortic aneurysms have an overall 5-year rupture risk of 30%.29 The Mount Sinai group has identified clinical variables that determine the risk for rupture, which include increasing age, presence of chronic obstructive pulmonary disease, maximal thoracic and abdominal aneurysm diameter, and the presence of pain.30 The Yale Aortic Diseases Group has documented rupture and dissection of ascending or arch aneurysms at a median size of 6 cm and descending or thoracoabdominal aneurysms at a median size of 7.2 cm.6 Furthermore, the Yale group has reported that the mean rate of rupture or dissection is 2% per year for small aneurysms, 3% for aneurysms 5.0 to 5.9 cm, and 6.9% for aneurysms of 6.0 cm and larger. Using proportional hazards regression, the odds ratio for rupture is more than 25 times higher in patients with aneurysms of 6.0 cm or greater than in those with aneurysms in the range of 4.0 to 4.9 cm.31

Open surgical graft replacement is the traditional treatment for these patients. The presence of comorbidities in this specific population increases the surgical risks, especially in the case of emergency intervention. However, with increasing experience, very good surgical morality rates in the 5 to 10% range have been reported from experienced centers.32–34 Similarly, paraplegia and paraparesis rates range from 3 to 16%, with predictive factors including extent of resection, emergency operation, renal dysfunction, distal circulatory support, and cerebrospinal fluid drainage.35–37 Five-year survival rates between 60 and 80% have been achieved in recent surgical series.32–34

Thoracic Aortic Dissections

Acute aortic dissection is the most common catastrophe affecting the thoracic aorta, with many more people dying of rupture of dissections than of aneurysms. Although poorly understood, a primary intimal tear allows a high-pressure entry of blood into the subadventitial space, which then rapidly propagates within the media proximally and distally. With the Stanford classification system, type A connotes involvement of the ascending aorta. The high mortality rate of 50% at 48 hours usually mandates emergent surgical repair. Conversely, type B dissections involve the descending thoracic aorta, and are typically managed with anti-impulse therapy, with surgical management reserved for those patients presenting with complications, namely intractable pain, rupture or impending rupture, or visceral or limb malperfusion syndromes.38–41

The Stanford group has compared the actual survival of medically and surgically treated type B dissection over a 36-year period. The actuarial survival estimates for all patients were 71, 60, 35, and 17% at 1, 5, 10, and 15 years, respectively, and were similar for the medical and surgical patients.40 The hope for benefits of avoiding the late complications from false-lumen expansion after surgical repair was not demonstrated in this follow-up.

Endograft repair of complicated acute type B aortic dissections is perhaps the greatest utility of this stent graft technology. Coverage of the primary intimal tear, redirecting flow into the collapsed true lumen, can dramatically reverse visceral malperfusion syndromes. The utility of thoracic endografts in chronic dissections is less clear. Given the multiple septal fenestrations, and the relative immobility of the chronically dissected septum, it seems unlikely that stent-graft insertion could realistically confer any long-term benefits, with the possible exception of a very focal aneurysmal false-lumen dilation distant from septal fenestrations near the level of the diaphragm.

Penetrating Atherosclerotic Ulcers and Intramural Hematomas of the Thoracic Aorta

Penetrating atherosclerotic ulcers (PAUs) and intramural hematomas (IMHs) are distinct pathologic entities now being diagnosed with increasing frequency.42 PAUs probably represent rupture of an atherosclerotic plaque, with penetration into the internal elastic lamina of the aorta, and may be associated with proximal and distal progression of an intramural hematoma. Conversely, IMH not associated with a penetrating ulcer may result from the spontaneous rupture of aortic vasa vasorum that may initiate hemorrhage into the aortic media, and may progress to an intimal tear and classic aortic dissection.42

In the ascending aorta, IMH, with or without PAU, frequently progresses to frank dissection during the acute phase, and thus warrants early ascending aortic replacement. The highest mortality rate among patients with IMH is associated with ascending aortic involvement.43,44 Experience has therefore suggested that a more aggressive approach with early surgery is warranted in those patients who have ascending aortic involvement or in those who have a coexisting aneurysm with IMH.44

In the descending thoracic aorta, pure IMH in the absence of aneurysmal change is usually treated with aggressive control of hypertension. For IMH with PAU, increasing maximal depth and maximal diameter were both associated with disease progression, in addition to persistent pain and increasing pleural effusion.45 The risk of aortic rupture is higher among patients with PAU by almost 30% than among patients with type A or B dissection.42 However, the progression of PAU is slow and is associated with a low incidence of acute rupture or other life-threatening events. Among patients with PAU who are not treated surgically, the natural history would indicate that the majority of patients will have aortic enlargement with the formation of saccular or fusiform pseudoaneurysms and intramural thrombus.43

Thoracic Aortic Trauma

Trauma is the most common cause of nondegenerative disease affecting the thoracic aorta. According to autopsy series, 36 to 54% of disruptions occur at the aortic isthmus, 8 to 27% involve the ascending aorta, 8 to 18% occur in the arch, and 11 to 21% involve the distal descending aorta.46 Blunt trauma is commonly a catastrophic injury with only approximately 20% surviving to hospital admission. Mortality after admission ranges from 39 to 73% and is frequently the result of other major injuries.47 The multiple injuries a patient may experience severely limit operative approaches and timing. All operations involving the thoracic aorta pose some risk of ischemic injury to the spinal cord, which is a dreaded complication in a predominantly young population. Closed-head injuries and other solid-organ injuries may constrain heparin use for peripheral circulatory support favored by most for open surgical repair.46 Fortunately, for these patients without radiographic signs of impending rupture, and with stable serial CT scans, permissive hypotension may be an effective temporizing strategy, allowing operative intervention after a patient has recovered from serious brain or lung injury that may have compromised immediate operative management.

Technical Development

The first stent grafts used at Stanford were manufactured using 2.5-cm self-expanding Gianturco Z stents (Cook Co., Bloomington, IN), which were fastened together and then covered with a woven Dacron graft (Meadox-Boston Scientific, Natick, MA; see Fig. 54-1). These stent grafts were oversized approximately 10 to 15% above the cross-sectional diameter ascertained by computed tomography (CT) in an effort to obtain sufficient radial force to achieve an endoseal and prevent stent-graft migration. A minimum of 2 cm of normal aorta was required for adequate fixation, otherwise referred to as the "landing zone," both proximally and distally. The covered stent was loaded into a delivery capsule that required femoral and iliac arteries greater than 8 mm to allow the introduction of a 28-French delivery sheath. This dilator contained a sheath that had been previously placed over a super-stiff guidewire and was positioned proximal to the point of deployment. Once this was achieved, the compressed stent graft was advanced into the sheath and deployed by using a "pusher" rod. Devices were limited to a maximal diameter of 40 mm in that aortas larger than 37 mm in diameter were themselves aneurysmal and unlikely to serve as stable attachment zones. Other anatomical constraints of these early "homemade" or first-generation stent grafts that precluded either delivery or secure fixation included acute angulation at the distal arch, and severe sigmoid-like tortuosity coursing through the diaphragmatic crura, reflecting the relative inflexibility of these early delivery systems.

The advent of this new stent-graft technology required a new terminology for endoleaks that allowed blood to leak around or through the stent graft, thus allowing the aneurysmal sac to remain pressurized. Type I endoleaks occur at the proximal or distal attachment sites, and signify a failure to achieve a hemostatic seal at these implantation sites.14,24,25 Type II endoleaks denote a communication between a branch vessel and the excluded aneurysm sac. These usually occur from a back-bleeding inferior mesenteric artery in the abdomen, or intercostal artery in the chest. Type III endoleaks originate from the middle graft sections, and are usually caused by disruption of graft-to-graft overlaps, or by leakage through the graft itself. Type IV endoleaks are characterized by an increase in size of the aneurysm sac in the absence of an identifiable patent branch vessel, variously referred to as endotension.

Years of experience with endovascular abdominal aorta repair and follow-up of thoracic aortic repairs has yielded important information for improved stent graft technology.11,48,49 Commercially produced second- and third-generation stent grafts are more flexible and have a lower profile, and thereby allow use of a smaller introducer sheath in the femoral vessels. Experience has shown that tapered, flexible, over-the-wire delivery systems that are less than 20 French in diameter rarely fail to traverse tortuous femoral or iliac arteries. Hooks at the proximal end of the stent graft appear to provide the most secure means of attachment, but may be suboptimal for treating patients with acute dissections. It is likely that different grafts may be developed for different pathologies, with devices for dissections being devoid of hooks and proximal uncovered metal components. For traumatic aortic lacerations, smaller device sizes are necessary, as these are usually nonatherosclerotic normal-sized aortas with small access vessels. Ideal device components have been broken down into three categories: delivery system, graft material, and metal frame.50 The delivery system should be of low profile, flexible for maneuverability, rigid enough to resist kinking, and hemostatic during use. The graft material should also be of low profile, strong and durable, and reasonably thin. Ideally, this material could also hold sutures. The graft metal frame should provide high column strength and ductility, be compression and kink resistant from external forces, radiopaque, and corrosion and fatigue resistant. Nitinol is now used for the stent material in the majority of grafts and the graft material is usually polytetrafluoroethylene (PTFE) or polyester.

Currently, the Gore Excluder TAG system (W.L. Gore, Sunnyvale, CA), the Medtronic Talent graft (Medtronic, Sunrise, FL), and the Cook Zenith (Cook Co., Bloomington, IN) are the only Food and Drug Administration (FDA)–approved thoracic grafts (Fig. 54-2). These second- and third-generation endoprostheses are more flexible and have a lower profile with smaller delivery systems that can be inserted easily to treat a number of thoracic aortic pathologies.

Clinical Results Using Endovascular Stents for Thoracic Aortic Disease

Thoracic Aortic Aneurysms

In January 2005 the phase II multicenter trial of the Gore Excluder TAG thoracic endoprosthesis results was reported.51 This multicenter prospective nonrandomized trial was conducted at 17 sites and compared results of stent-graft repair of descending thoracic aortic aneurysms in 140 patients with results of open repair in 94 patients. Strict inclusion and exclusion criteria were defined in an attempt to ensure comparability of both groups. Follow-up CT scans were obtained at 1, 6, 12, and 24 months. For stent-graft patients, operative blood loss, renal failure, paraplegia, and mortality rates were all significantly less than for the open repair group (Fig. 54-3). Interestingly, stroke rates were about equal in both groups. ICU stay and total hospital stay, and time to return to normal activity were 50% shorter for the stent-graft group than for those with open repair. Although stent-graft patients maintained an advantage from aneurysm-related mortality out to 2 years (97 versus 90%), interestingly, all-cause mortality was similar between groups at 2 years, which is similar to results of recent randomized trials in abdominal aortic aneurysm stent-graft trials.52

Ricco and colleagues have most recently reported on an independent nationwide study in France using a variety of endovascular devices to treat descending thoracic aortic aneurysms in the majority of cases.53 The Gore Excluder TAG (Gore) and Talent (Medtronic) devices were used in 84% of patients and an operative mortality of 10% was reported. A complication rate of 21% was reported, which included endoleaks in 16% of patients that were fatal in three patients. The 6-month survival rate was 86% and freedom from other complications (other than endoleak) was only 63% at 6 months. This study, involving 166 stent-graft repairs, performed in 29 centers and using six different types of endoprostheses, demonstrated that stent-graft repair of thoracic aortic disease could be performed with acceptable morbidity and mortality, at short-term follow-up. They do, however, qualify their conclusions by stating that endograft treatment of thoracic aortic aneurysms should continue to be used in an investigative setting.

Perhaps similar to abdominal aortic aneurysms, stent grafts may become the preferred treatment for ruptured thoracic aortic aneurysms. Previously, open surgical repair of ruptured thoracic aortic aneurysms was associated with significant mortality and morbidity. Specialized centers with both expertise and appropriate devices may be capable of endovascular repair in patients with sufficient 2-cm landing zones to allow secure fixation of an appropriately sized endograft. Although thoracic ruptures are far less common than abdominal aneurysm ruptures, the increasing penetrance of this technology will likely allow timely intervention in multiple specialized centers.

Thoracic Aortic Dissection

For dissections originating distal to the origin of the left subclavian artery, referred to as Stanford type B dissections, aggressive antihypertensive therapy has been the mainstay of treatment.40 For complicated dissections, however, defined as those dissections with rupture, impending rupture, intractable pain, rapid expansion, or malperfusion syndromes, surgical intervention is indicated. In these instances, however, surgical mortality has been reported to be as high as 50 to 60% in this high-risk group of patients. Endograft coverage of the primary intimal tear, redirecting flow into the true lumen, appears to be an ideal application of this endograft technology. The Stanford group, with their colleagues in Mie University in Japan, initially reported the use of the first-generation stent grafts in 19 patients with complicated type B aortic dissections. Placement was successful in all patients, and revascularization of ischemic branch vessels occurred in 76% of cases. There were three hospital deaths, two of which resulted from late referral, with irreversible end-organ damage, and one in a patient with Ehlers-Danlos syndrome who was perhaps untreatable by any modality. Although the follow-up was short, there were no incidences of aortic rupture or aneurysm formation, and a single-lumen aorta to the level of the stent graft was achieved in the majority of patients27 (Fig. 54-4).

Multiple groups have now reported their results with endograft repair of complicated acute type B aortic dissections. Dynamic malperfusion is typically reversed with stent-graft coverage of the primary intimal tear, restoring flow into the true lumen and expanding collapsed segments. Static mechanisms of malperfusion, with branch vessel occlusion from orifice tears of the intima, can be diagnosed during the same intervention, and frequently reversed with uncovered stents into the true lumen of branch vessels. Mortality rates for this high-risk population have been significantly lowered from the previous surgical figures of 30 to 60%.

Likely because of these good results, interest developed in the utility of stent grafts for uncomplicated acute type B aortic dissections. A randomized trial referred to as the INSTEAD trial (INvestigation of STEnt in patients with type B Aortic Dissection)* investigated all-cause mortality as the primary outcome; secondary outcome variables included conversion to stent and/or surgery, thrombosis of the false lumen, cardiovascular morbidity, aortic expansion, quality of life, and hospital stay. Given the relatively low 1-year mortality for uncomplicated type B dissections, especially if treatment is delayed for at least 2 weeks, it is not surprising that the survival of the stent-graft population was inferior to that achieved with standard medical therapy. Another randomized multicenter trial, the ABSORB trial, will randomize uncomplicated patients during the acute phase, and is currently enrolling patients.54 The early experience suggests that, in experienced hands, stent-graft morbidity is low. Covered portions should not extend below T-6 to T-7 to minimize the incidence of paraplegia. Proximal uncovered stents should not be placed in the curvature of the aortic arch to avoid retrograde extension into a more dangerous type A dissection. More distally, uncovered stents may promote false-lumen thrombosis while allowing continued perfusion of intercostal arteries. False-lumen thrombosis is likely for the extent of stent coverage. Stent grafts specifically designed for dissections may well incorporate these features, ie, flexible stent grafts with low hoop strength, no uncovered elements on the proximal end, approximately 10 cm of covered stent, and a longer segment of uncovered stents distally.

As stated, the utility of stent grafts for the management of chronic aortic dissections is quite problematic. Because of disease complexity and extent, many were hopeful that stent grafts would prove an effective modality for treatment. However, the increasing thickness of the dissection flap with time, the frequent severe narrowing of the true lumen, the variable origin of visceral vessel origins from both true and false lumens, and the presence of multiple fenestrations between true and false lumens have severely limited the utility of stent grafts in the management of chronic aortic dissections. One exception may be the focal enlargement of the false-lumen in the proximal descending thoracic aorta. Coverage of the proximal intimal tear in these patients may allow false-lumen thrombosis for the proximal thoracic aorta. However, fenestrations almost invariably present at the level of the diaphragm and below allows a high- pressure entry into the false lumen, which then frequently propagates proximally. Although some local growth may be limited, further enlargement usually continues distally beyond the extent of the stent graft. More distal coverage in an attempt to promote false-lumen thrombosis may increase the risk for paraplegia. Uncovered stents may stabilize the flap, promote false-lumen thrombosis, and allow continue perfusion of patent intercostals arteries. These strategies will require further investigation.

Penetrating Atherosclerotic Ulcers and Intramural Hematomas

IMH of the aorta is attracting growing interest as a variant of aortic dissection.44 The exact pathophysiology is not well understood. Although by definition, pure IMH likely occurs from hemorrhage into the media from the vasa vasorum, many maintain that an intimal disruption is present in all cases.

Certainly, in the absence of any intimal disruption, there would be no indication for stent-graft repair.44 IMH, however, is often associated with or even precipitated by PAUs of the descending thoracic aorta.42 Therefore, covering the PAU with a stent graft may limit the progression of the IMH and allow healing to occur.19,42 Unfortunately, even with successful stent graft implantation using both first- and second-generation grafts, retrograde aortic dissection, new ulcer formation, and endoleaks have been noted in a significant percentage of patients, emphasizing the diffuse and severe nature of this disease.55–58

The Stanford group has reported their midterm results treating PAU of the descending thoracic aorta, with an average of 51 months of follow-up19 (Fig. 54-5). Using both first- and second-generation commercial devices, 26 patients were treated, 14 of whom were deemed nonoperative candidates. The primary success rate was 92%, with actuarial survival estimates of 85, 76, and 70% at 1, 3, and 5 years, respectively. Perioperative mortality was 12%. Increasing aortic diameter and female gender were determinants of treatment failure. These risk factors reflect the importance of careful patient selection based on anatomical criteria and clinical factors. In addition, long-term follow-up with serial CT angiography is necessary to detect late complications.

Thoracic Aortic Trauma

Aortic injuries secondary to nonpenetrating trauma are lethal lesions, with 80 to 90% of patients dying in the hour after the accident.59 Urgent surgical graft replacement of the aorta has been the standard treatment, but these patients frequently have other major injuries, including closed-head injuries, pulmonary contusions, and other solid-organ injuries, which may limit options for open surgical repair. Several authors have reported the improved results of stent grafts over open repair for these acute injuries.60,61 Although long-term durability may be a concern, there appear to be significant short-term benefits.61,62 The major difficulty at present is the absence of thoracic stent grafts sufficiently small in diameter as to be appropriate for these relatively normal-sized aortas, frequently less than 20 to 22 mm in diameter. Additionally, small iliac and femoral vessels may limit access. Because of these limitations, and perhaps because we rarely see patients within the first few hours of their injury, during which the risk for late rupture is probably greatest, we have embarked on a strategy of permissive hypotension, intervening only for those patients with signs of impending rupture, those with increasing mediastinal hematoma or hemothorax, or persistent pain. For these patients, conventional repair is preferred, using heparin-bonded circuits, unless closed-head injury or pulmonary contusions contradict. Stent-graft repair is used for those patients in whom a conventional repair could not be tolerated. Computed tomography is repeated at 24 hours, and again intervention is elected only if there has been enlargement or progression in the extent of the pseudoaneurysm. Under very careful surveillance, either open or endovascular repair is then performed when the patient has sufficiently recovered from his or her other injuries. It is likely that, as stent grafts with more suitable characteristics become available, they may become the repair of choice.

There are no reliable reports of sufficient numbers to afford real long-term assessment of stent-graft performance. Modern devices appear to have good 5-year durability, although reports of fabric tears associated with sharp angulations or grafts within grafts still surface occasionally. More important, even with complete exclusion of the aneurysm sac, aortic elongation may still occur, allowing the late development of type I endoleaks. Therefore, lifelong monitoring is essential, and currently requires cross-sectional imaging with either computed tomography or magnetic resonance imaging. For these reasons, we have preferred open repair for younger good-risk patients, using endovascular repair for older individuals with favorable anatomy. The Stanford group has reported on their midterm results of stent-graft repair of chronic traumatic aneurysm of the descending thoracic aorta63 (Fig. 54-6). Among 15 patients treated with either first- or second-generation stent grafts, deployment was successful in all patients without need for surgical conversion. No neurologic complications were reported. Actuarial survival estimates at 1 and 6 years were 93 and 85%, respectively. Freedom from reintervention on the descending thoracic aorta was 93 and 70% at 1 and 6 years, respectively. Freedom from treatment failure at 1 and 6 years was 87 and 51%, respectively. They therefore concluded that stent grafts are safe in patients with chronic traumatic aneurysms, and are associated with satisfactory but not optimal midterm durability. They state that younger, low-risk patients should be offered conventional, open surgery and stent-grafting should be reserved for those patients who are at prohibitive operative risk.63

*Nienaber C, Rousseau H, Eggebrecht H, et al: A randomized comparison of strategies for type B aortic dissection: the Investigation of STEnt grafts in Aortic Dissection (INSTEAD) trial. Circulation 2009; 120:2513-2514.

Despite the many advances in the field of endovascular surgery, stent-grafting the thoracic aorta remains in a developmental phase.20,64 This evolving technology has been applied to the treatment of thoracic aortic aneurysms, aortic dissections, IMH, PAUs, and traumatic injuries. The early outcomes are encouraging but middle- and long-term outcomes are a concern. Although device technology has certainly improved over the past 20 years, all devices are limited by their own structural flaws, lack of conformability, limited array of sizes, length of landing zones necessary to allow secure fixation, and structural integrity to withstand the severe physiologic milieu present in the thoracic aorta. Given the variable pathologies, it is likely that many different stent-graft configurations will be necessary for the various clinical indications.

Of utmost importance in the treatment of thoracic aortic pathology with stent grafts is strict and dedicated follow-up. Patients need to be seen on a routine basis and new symptoms or findings investigated. Serial CT angiography is an excellent tool to follow the areas of the thoracic aorta treated with the endograft, as well as to follow the evolution of the untreated portions of the aorta. Follow-up will allow endoleaks and device migration to be detected, and may further define the natural history of endovascular treatment strategies.

Diseases of the thoracic aorta pose a significant challenge to the surgeon because of the complexity of the disease and the characteristics of the patient population. Current literature supports the early advantage of endovascular technology in well-suited patients, but longer follow-up and results of ongoing trials will help define the indications for their use in the future.