Thoracic aortic lesions treated with the Zenith TX1 and TX2 thoracic devices: Intermediate- and long-term outcomes
Article Outline
Objective
Little data exist to support the durability of thoracic endovascular repair during prolonged periods of follow-up. This study examines the durability and long-term results with the Zenith TX1 and TX2 thoracic devices (Cook Inc, Bloomington, Ind) in high-risk patients.
Methods
Data were collected prospectively from 2001 to 2007 on high-risk patients who presented with thoracic aneurysms, chronic aortic dissection, or fistulas treated with a Zenith thoracic device. Surgical modifications of proximal or distal landing zones were performed when necessary. Computed tomography follow-up scans were performed before discharge, at 1, 6, and 12 months, and yearly thereafter. Three-dimensional reconstruction software with central line of flow measurements was used to assess aortic morphologic characteristics. Kaplan-Meier analysis was used to assess survival, freedom from reintervention, predictive factors of poor outcome, and morphologic changes, including aneurysm sac behavior.
Results
A total of 160 patients (44% women; mean age, 70) were treated for 130 thoracic aneurysms, 25 aortic dissections with aneurysm, 2 fistulas, and 3 symptomatic or aortic ruptures, or both. Mean follow-up was 36 months, and aneurysm size was 67 mm. Seventy-five patients (47%) had undergone prior aortic aneurysm repair. Surgical modifications were required to create adequate landing zones in 33% patients, including 28 elephant trunk/arch reconstruction, 22 carotid-subclavian bypasses, and seven visceral vessel bypasses. Iliac conduits were required in 31 patients. Early mortality (<30 days) occurred in 11 patients (6.9%). Overall mortality at 1 year was 16%. Aneurysm sac increase (>5 mm) requiring intervention was observed only in one patient in the settings of component separation and type III endoleak that was treated; the sac is now stable. Twenty-seven endoleaks were detected in 25 patients: 15 primary endoleaks (9.4%) <30 days and 12 secondary endoleaks (7.5%) >30 days. Secondary interventions were required in 42 patients (26%).
Conclusion
Endovascular treatment of thoracic aortic pathologies with the Zenith TX1 and TX2 devices is feasible and durable. The mid- to long-term results are encouraging, with acceptable low reintervention rates and with good survival within high-risk patients.
The use of endovascular stent grafts has revolutionized the management of thoracic aortic pathologies. This modality is considered safe and effective at preventing aneurysm rupture through short- and intermediate-term follow-up studies; however, little extended follow-up data exist in the published literature.1, 2, 3, 4, 5, 6 The ascribed benefits of such treatment options center on the lack of need for a thoracotomy, extensive tissue dissection, and aortic cross-clamping. The acute physiologic insult is less compared with conventional open surgical repairs, allowing physicians to treat patients considered too ill for traditional surgery and lessening the risk of morbidity and mortality for those who are candidates for either type of repair.7
Although some randomized trials support a long-term benefit for infrarenal aneurysm repair,8 similar studies have not been conducted with thoracic aortic pathology. The less frequent appearance of thoracic disease compared with infrarenal aneurysms offers some explanation; however, thoracic aortic disease may arise from a variety of etiologies, each having marked differences in outcome expectations (diverticulum, aneurysms, and aortic dissection). These factors confound the ability to design, analyze, and report outcomes for thoracic endovascular aneurysm repair (TEVAR).
All published studies either represent series of treated patients or controlled trials where endovascular patients are compared with a mixture of retrospective and prospective surgical patients serving as a control population. Despite the lack of optimal data, insight with respect to the benefits and challenges of endovascular repair can be gained from a detailed evaluation of clinical experiences. Concerns specific to the durability of endovascular repair include device migration, late endoleak occurrence, component separation, and device integrity, all of which can only be assessed in the setting of detailed and extensive clinical and radiographic follow-up. The aim of this analysis is to examine the extended follow-up of TEVAR using the Zenith TX1 and TX2 stent grafts (Cook Inc, Bloomington, Ind) in high-risk patients with descending thoracic aortic pathology.
Methods
Data were accumulated under the guidance of a sponsored investigational device exemption that began in February 2001 and ended in August 2007 (for the purpose of this analysis). Enrolled patients were considered to be at high risk for open surgical repair and signed an informed consent form approved by the Cleveland Clinic Institutional Review Board. Inclusion criteria and procedural details have been previously published.4
Adherence with the latest version of the endovascular grafting reporting standards was attempted whenever possible, with the notable exception of device migration, which was analyzed in a manner previously described by our group.9 All imaging data are reported using three-dimensional (3D) analysis tools (multiplanar reconstructions, centerline of flow calculations, and surface rendering algorithms) when appropriate. Paraplegia was defined as a complete deficit or inability to ambulate as a result of spinal cord ischemia at the time of hospital discharge or death. Operative time was defined as the time when the skin incision was made to skin closure.
The device design has also been previously described; however, two fundamental delivery system changes were made during patient enrollment. The first alteration was implemented to diminish sheath kinking. The standard polytetrafluoroethylene sheath that was used on early devices was replaced with Flexor material (Cook Inc), which is a braided sheath with hydrophilic coating. Although this improved some delivery characteristics, the rigidity of the system remained a challenge. Given that the device is delivered by insertion over a stiff wire, it would be optimal to match the device flexibility with that of the insertion wire. To accomplish this, the inner cannula through which the stiff wire passes was changed from a tube of stainless steel to a nitinol cannula. This change markedly improved the flexibility and was termed “super-flex,” but later renamed to Z-Track Plus.
Follow-up and statistical analysis
Clinical, radiographic, and laboratory follow-up were obtained before discharge, at 1, 6, and 12 months, and yearly thereafter. Mortality data was supplemented by querying the Social Security Death Index quarterly and follow-up phone calls when necessary. Data were entered into an Oracle Clinical database (Oracle Corp, Redwood Shores, Calif) for the purposes of compliance and edit tracking. Kaplan-Meier techniques, univariable, multivariable analysis, and logistic regression were used in an effort to relate outcomes with preoperative factors. Statistical analyses were performed with SAS software (SAS Institute, Cary, NC).
Results
A total of 160 patients (44% women), with a mean age of 70 years, were treated. Indications for intervention according to the disease etiology included thoracic aneurysms in 130, chronic aortic dissection with aneurysm in 25, aortobronchial fistulas in 2, and symptomatic or aortic ruptures, or both, in 3. Mean follow-up was 36 months (range, 1 day-78 months), and mean aneurysm size was 67 mm. The average length of aorta covered by the stent graft, as measured by a centerline of flow analysis, was 219 mm. Despite a liberal minimum proximal and distal neck length of >10 mm, 33% of patients required surgical modification of the implantation landing site. Elephant trunk grafts were placed in 28 patients, 22 had preoperative carotid–subclavian bypasses, and seven required visceral bypasses. Two patients who required elephant trunk grafts also required carotid–subclavian bypass, and two underwent visceral vessel bypass as well. Iliac conduits for access were required in 31 patients (19%). At least one prior aortic repair had already been done in 75 patients (47%) in this cohort (Table I).
Table I. Demographics and aneurysm type
| Variablea | Thoracic aneurysm (n = 131) | Chronic dissection aneurysm (n = 25) | Other (n = 4) | Total (n = 160) |
|---|---|---|---|---|
| Age, yearsb | 74 (68-79) | 55(46-71) | 74(62-80) | 73(66-79) |
| Menb | 68(52) | 17(68) | 4(100) | 89(56) |
| Prior aortic repairb | 62(47) | 12(48) | 1(25) | 75(47) |
| Aneurysm size, mean mmb | 66(58-76) | 59(54-68) | 75(62-82) | 65(57-75) |
| Surgical modification of the proximal neck | 39(30) | 8(32) | 1(25) | 48(30) |
| Visceral bypass | 7(5) | 0 | 0 | 7(4) |
| Iliac or aortic conduit | 29(22) | 2(8) | 0 | 31(19) |
| Anesthesia | ||||
| General | 41(31) | 9(36) | 4(100) | 54(34) |
| Regional | 88(67) | 15(60) | 0 | 103(64) |
| Local | 2(2) | 1(4) | 0 | 3(2) |
| Surgical time, min | 142(96-187) | 115(95-195) | 162(139-178) | 140(96-190) |
| Contrast volume, mL | 135(105-185) | 132(90-170) | 175(132-190) | 137(105-185) |
| Fluoroscopy total time, min | 16(12-24) | 15(9-18) | 19(17-25) | 16(12-23) |
| Estimated blood loss, mLb | 300(200-600) | 200(100-350) | 225(131-350) | 300(195-562) |
aContinuous variables are given as median (interquartile range, 25th-75th percentile) and categoric variables as number (%). |
bP < .05 for comparison among groups; caution with small numbers. |
Mortality
Early mortality (<30 days) occurred in 11 patients (6.9%). Two patients each died of primary cardiac issues, including ventricular tachycardia in one and myocardial infarct in one, and cerebrovascular accident (CVA). Two patients died of bleeding, one resulting from continued bleeding from an aortoesophageal fistula, which was the primary indication for TEVAR; in the second patient, a coagulopathy developed relating to a ruptured iliac artery during delivery complications. Two died of sepsis (related to a perforated diverticulum in 1 and as a result of uncontrolled fungemia in 1), and one patient each died of pulmonary edema, multisystem organ failure, and an undefined cause. Overall survival was 84%, 78%, 73%, 70%, and 70% at 12, 24, 36, 48, and 60 months, respectively (Fig 1 and Table II).
Fig 1.
Overall mortality (thick line) and aneurysm-related morality (thin line) were assessed using a Kaplan-Meier analysis, and the error bars denote the 95% confidence intervals (CI).
Table II. Outcomes of treatment classified in terms of presenting pathology
| Outcome at follow-up | Overall (n = 160) | Thoracic (n = 131) | Chronic dissections (n = 25) | Other (n = 4) | |
|---|---|---|---|---|---|
| A (n = 104) | B (n = 27) | ||||
| Survival, % | |||||
| 1 year | 84 | 83 | 82 | 92 | 75 |
| 2 years | 78 | 77 | 77 | 88 | 75 |
| 3 years | 73 | 68 | 77 | 88 | 75 |
| 4 years | 70 | 64 | 77 | 88 | 75 |
| 5 years | 70 | 64 | 77 | 88 | 75 |
| Free of 2nd interventions | |||||
| 1 year | 81 | 78 | 81 | 88 | 75 |
| 2 years | 76 | 76 | 70 | 83 | 75 |
| 3 years | 74 | 74 | 70 | 79 | 75 |
| 4 years | 73 | 73 | 70 | 79 | 75 |
| 5 years | 65 | 63 | 70 | 79 | 75 |
| Free of migration | |||||
| 1 year | 99 | 99 | 96 | 100 | 100 |
| 2 years | 99 | 99 | 96 | 100 | 100 |
| 3 years | 94 | 96 | 82 | 100 | 100 |
| 4 years | 93 | 94 | 82 | 100 | 100 |
| 5 years | 93 | 94 | 82 | 100 | 100 |
Neurologic complications
Paraplegia occurred in five patients (3.1%), three of which died during the perioperative period before any significant neurologic recovery. A single patient recovered completely ≤6 months, and the remaining patient had a persistent unilateral lower extremity deficit. All of the patients who presented a neurologic deficit had extensive aortic coverage. The mean (SD) length of the aorta covered in patients who had a spinal cord ischemic insult was markedly greater then the length of covered aorta in patients without deficits, at 219 (68.45) mm vs 315 (44.09) mm (P < .001). Three of the five patients also had complications such as iliac rupture in the setting of Ehlers-Danlos disease, sepsis, or renal failure. Associated risk factors observed in patients with paraplegia included single internal iliac arteries and prior abdominal aortic aneurysm (AAA) repair in four of the five patients, perioperative hypotension in two, and left subclavian artery coverage in two, although both had undergone carotid subclavian bypass before TEVAR.
During the perioperative period, five patients (3.1%) presented with CVAs and four died. Two patients died early, at 10 and 25 days, and two others died after an additional stroke at 93 and 375 days. Of interest, four of the five CVAs involved the posterior circulation and one the frontoparietal cortex. Three of the five patients had extensive aneurysmal disease of the aortic arch and underwent staged carotid–subclavian bypasses. The single long-term survivor of a CVA had presented with a ruptured thoracic aneurysm and was treated with an urgent carotid–subclavian bypass graft, followed by stent graft placement at the same setting. This patient sustained a posterior circulation stroke, recovered significantly, and was ultimately discharged with warfarin therapy. The two other patients who survived their initial strokes (1 posterior circulation, 1 frontoparietal cortex) sustained neurologic deficits during balloon inflation in the arch while undergoing TEVAR with regional anesthesia. Both recovered partially, had prior carotid–subclavian bypasses, and were discharged with aspirin therapy. One died as a result of a second stroke that affected the same cerebrovascular distribution, and the other died of a ruptured subclavian aneurysm in the setting of a properly placed endograft, as determined by autopsy.
Secondary procedures
Secondary interventions were required in 42 patients (26%). Indications for reintervention specifically related to long-term issues with the treatment or device include the 9 endoleaks, 2 aortobronchial fistulae (1 was a recurrence of an aortobronchial fistula that represented the initial treatment indication; the second occurred de novo during follow-up), 1 aneurysm sac growth, 3 component separations, 1 compressed stent within an elephant trunk graft, and 1 patient with what was perceived to be a tenuous proximal sealing segment in the absence of an endoleak. Endovascular techniques were used to treat all but one patient. In the surgically treated patient, who had a persistent type I proximal endoleak, we placed an elephant trunk graft proximally and combined the elephant trunk graft with the endovascular repair distally using a second thoracic component. This was accomplished successfully. The breakdown of secondary interventions and life-table analysis of freedom from requiring a secondary intervention are presented in Fig 2 and Table II, Table III.
Fig 2.
Freedom from secondary intervention is shown in 160 patients with >72 months of follow-up.
Table III. Description, type, and indications for 42 secondary interventions
| Cause | No. | Open | EVAR | % |
|---|---|---|---|---|
| Related to DTA repair | 19 | 45 | ||
| Endoleak | 9 | 1 | 8 | 21 |
| Fistulaa | 2 | … | 2 | 5 |
| Aortic growth | 1 | … | 1 | 2 |
| Proximal dissectionb | 1 | … | 1 | 2 |
| Component-separation | 3 | … | 3 | 7 |
| Compressed-stent | 1 | … | 1 | 2 |
| LCC-LSA bypass | 2 | 2 | … | 5 |
| Inadequate seal | 1 | … | 1 | 2 |
| Procedure related | 9 | 21 | ||
| Thrombectomy | 3 | 3 | … | 7 |
| Hematoma evacuation | 1 | 1 | … | 2 |
| Leg ischemia | 2 | … | 2 | 5 |
| Brachial pseudoaneurysm | 2 | … | 2 | 5 |
| Rupture Iliac | 1 | 1 | … | 2 |
| Associated aneurysms/operations | 14 | 21 | ||
| AAA | 7 | 1 | 6 | 17 |
| Pseudoaneurysm | 2 | … | 2 | 5 |
| Ascending repair | 2 | 2 | … | 5 |
| Thoracoabdominal | 1 | … | 1 | 2 |
| Iliac | 1 | 1 | … | 2 |
| Carotid endarterectomy | 1 | 1 | … | 2 |
| Total | 42 | 13 | 32 | 100 |
aIn the two patients with fistulas, one had a recurrent aortoesophageal fistula and the other had a de novo aortoesophageal fistula that was not seen on angiogram but autopsy report revealed a fistulous track between esophagus and the aorta just above the proximal stent graft. |
bThis patient presented with a type A dissection before the endovascular repair, underwent treatment of the distal portion first, and subsequently was treated for the proximal dissection. |
Sac size and endoleaks
Aneurysm sac increase (>5 mm) was observed in only two patients (Table IV). One patient had aneurysm growth noted at the 1-year follow-up visit, without evidence of endoleak or compromised fixation. She was monitored, and the aneurysm had decreased in size at her 2-year visit. She is now 3 years after treatment, without evidence of sac expansion. Only one other patient had sac size increase that was detected at 4 years in the setting of a type III endoleak resulting from component separation. The patient underwent placement of an additional component to eliminate the endoleak and has not had further growth to date. Primary endoleaks (<30 days) occurred in 15 patients (9.4%), and late endoleaks (>30 days) were noted in 12 patients (7.5%). Other endoleaks detected were 10 type I (7 primary, 3 secondary), 14 type II (7 primary and 7 secondary), as well as 3 type III (1 primary and 2 secondary). Two patients had endoleaks with two separate causes. The first patient had a primary type I endoleak, which was treated with a secondary intervention perioperatively, and then presented with a late type II endoleak, which has been observed. The second patient presented a late type II leak that was observed until component separation and aneurysm growth were noted, thus prompting a secondary intervention. Freedom from endoleaks is shown in Fig 3.
Table IV. Aneurysm sac behavior during follow-up period
| Sac behavior | 1 month (n = 130) | 6 months (n = 109) | 1 year (n = 106) | 2 years (n = 82) | 3 years (n = 48) | 4 years (n = 24) | 5 years (n = 18) |
|---|---|---|---|---|---|---|---|
| Sac growth, % (No.)a | 0%(0) | 0%(0) | 1%(1) | 0%(0) | 0%(0) | 4%(1) | 0%(0) |
| Sac stable, % (No.) | 65%(84) | 43%(47) | 31%(33) | 28%(23) | 27%(13) | 38%(9) | 17%(3) |
| Sac shrink, % (No.) | 35%(45) | 57%(62) | 68%(72) | 72%(59) | 73%(35) | 58%(14) | 83%(15) |
aNumbers in parenthesis reflect number of patients with a computed tomography scan in that follow-up period |
Fig 3.
A, Kaplan-Meier life-table analysis shows of freedom from primary endoleak and (B) secondary endoleaks.
Migration
Migration was assessed by centerline of flow distances from fixed landmarks (left common carotid and celiac arteries), where 14 patients had >10 mm of distance change throughout follow-up. Any patient with detected changes of baseline central line length measurements underwent further image assessment, primarily using surface-rendering reconstruction to evaluate device position in relation to local landmarks, such as hemaclips placed on elephant trunk grafts or focal aortic calcifications.9 These analyses showed device movement (Fig 4) had occurred in four patients (2.5%). The details of all migration patients are listed in Table II, Table V.
Fig 4.
A, A three-dimensional (3D) reconstruction of a computed tomography (CT) scan demonstrates an endovascular completion of an elephant trunk graft. The pacer wire attached near the distal end of the elephant trunk graft is indicated by the small arrow. Distally, the device was deployed into an aneurysmal segment that had been wrapped with a polyester strip to create a landing zone in a patient deemed unable to tolerate an open thoracoabdominal repair (large arrow). B, Same patient, 3D reconstruction CT scan at the 6-month follow-up demonstrates a stable proximal dense position (arrow) enlargement of the wrapped segment, distal endoleak, and proximal migration (circle) of the distal stent through the wrap and into the aneurysm. Note, no barbs were used on the distal fixation system for fear of injury to the abnormal aorta below the level of the stent graft. C, Magnified view of previous image demonstrates scratches on the external aortic wall (arrows) caused by migration of the device.
Table V. Patients diagnosed with device migration based on central line of flow and surface rendered analyses
| Date of event | Stent | Secondary intervention | Outcome | ||
|---|---|---|---|---|---|
| No. | Type | Migrated | |||
| 2 years | 1 | TX1 | Distal | Yes | OK |
| 3 years | 1 | TX1 | Proximal | Yes | OK |
| 1 year | 1 | TX1 | Proximal | No | Withdrew from study |
| 1/2 year | 3 | TX2 | Distal | Pending | Pending |
Device integrity
Barb fractures occurred in eight patients (5%). Depending on device diameter, 12 or 13 barbs exist proximally and distally, and all patients with noted barb fractures had only a single barb break (Fig 5). Fractures were detected in three patients at 24 months, in two at 36 months, and in three at 60 months. Only one of the eight patients with barb fractures had evidence of device migration, which was noted at 2 years of follow-up and treated with a proximal extension (Table VI). The remaining seven patients with barb fracture occurred in the setting of stable fixation and sealing zones.
Fig 5.
A and B, Radiographs in different projections demonstrate a single barb fracture (circle and arrow) first noticed at 2 years follow up.
Table VI. Patients identified with single barb fractures along with the type of device, timing, and association with any clinical events
| Date of event | Stent | Barb fractures, No. | Migration | |
|---|---|---|---|---|
| No. | Type | |||
| 5 years | 1 | TX1 | 1 | … |
| 5 years | 1 | TX1 | 1 | Yes |
| 3 years | 1 | TX1 | 1 | … |
| 5 years | 1 | TX1 | 1 | … |
| 3 year | 2 | TX1 | 1 | … |
| 2 years | 2 | TX2 | 1 | … |
| 2 years | 2 | TX2 | 1 | … |
| 2 years | 2 | TX2 | 1 | … |
Patient accountability
Mortality data were available for all patients throughout follow-up, as described in Methods. Clinical and imaging data were available for most patients: 86% of patients had imaging and clinical information available for analysis at 2 years, and 73% of the patients had complete imaging and clinical data as well at 5 years.
Discussion
Descending thoracic aortic aneurysms (TAA) are potentially life-threatening diseases. The incidence and prevalence of degenerative TAA have increased significantly,10, 11, 12, 13 likely as a result of increased awareness, screening, and an aging population. The effect of this phenomenon will be an increasing need for TAA intervention. Conventional open resection and graft replacement of the pathologically altered aorta has been considered to be the standard of care.14, 15 However, even in centers of excellence where great strides have been achieved with respect to the development of new surgical techniques and methods of patient management, the morbidity and mortality rates for open TAA repair remain high.16, 17, 18, 19 A considerable population of patients with thoracic aortic pathology are refused open surgical repair as a result of comorbid illnesses and ultimately die of TAA rupture.
TEVAR has emerged as a treatment option for patients amenable to open repair and those considered too ill for open repair.4, 5, 20, 21, 22, 23 Several published studies have followed the initial published report by Dake et al3 in 1994. Despite some study design weaknesses, the first completed comparative multicenter trial demonstrated convincing evidence of superiority of TEVAR vs open surgery with respect to both survival and morbidity.7, 24 Yet, the longevity of the endovascular repair (migration of device and endoleaks) remains controversial, may depend on the etiology of the disease and the specific portion of the anatomy treated, and requires further elucidation.
The etiology of the disease, extensiveness of the involved aorta, patient selection, and procedural planning are paramount to the success of TEVAR. Simple vs complex anatomy and straight healthy landing zones vs tortuous diseased aorta all have marked implications on the expected outcome. Unquestionably, endovascular procedures are less invasive than open surgical TAA repair; yet even TEVAR is associated with a relatively high perioperative mortality rate. Despite stratification of a variety of comorbidities, true predictors of poor outcome were not possible in this series given the limited number of patients and the diversity of preoperative risk factors, including age, TAA size and etiology, pulmonary dysfunction, cardiac disease, staged hybrid procedures, and chronic renal insufficiency.
These issues bespeak to the importance of clinical judgment required to have a successful TEVAR program. A number of patients were evaluated for treatment and not considered to be candidates for endovascular repair because of physiologic or anatomic challenges. The outcome of such patients is the subject of a separate study. In contrast with other series, our clinical protocol did allow for the treatment of patients with very short and otherwise challenging landing zones, markedly tortuous anatomy, and aneurysms elsewhere in the aorta, as well as the liberal use of iliac/aortic conduits. Hybrid procedures, such as the placement of elephant trunk grafts,25 were frequently used in preference to reliance on a compromised sealing or fixation region but were not possible in all patients. Retrospective analysis of preoperative imaging studies of the four patients noted with device migration demonstrated short, tapered fixation zones, implying that device placement was planned into unhealthy aorta. These regions demonstrated evidence of aortic dilation, likely resulting in inadequate fixation strength that most probably contributed to the observed migration.
When patient were amenable to additional procedures, migration cases were treated with secondary interventions intended to bring the fixation zone into healthy aorta by using either fenestrated devices or extra-anatomic bypasses, followed by endograft extensions. Obviously, to prevent this, we would recommend planning the procedure such that the proximal and distal fixation and sealing zones exist in healthy aorta. Yet, the assessment of the health of the aorta is far from a pure science.
We currently rely on the appearance of the intended fixation zones on a stretched centerline of flow reconstruction or a curved planar reconstruction. Tapering diameters, nonparallel walls, and debris or thrombus lining the wall are generally considered to be unhealthy. These observations were exceptionally difficult to make on 2D imaging studies and really require 3D image manipulation for proper assessment. Ultimately, it is likely that the choice of landing zones is the most important component of procedural planning. In addition to the status of the aortic wall, the location with respect to tortuosity and aortic branches must be considered. After the establishment of a plan for fixation/sealing regions, the length of the intervening segment was used to dictate the size and number of components required to complete the repair, a factor that had implications for the risk of neurologic complications.
Catastrophic neurologic complications, paraplegia and strokes, have been reported in up to 18% patients undergoing TEVAR.26 Although the incidence of spinal cord ischemia (SCI) for thoracic stent grafts may compare favorably with open surgical repairs, it is likely that the patient populations differ, rendering firm conclusions suspect to bias. Obviously, it is not possible to reimplant intercostal arteries during TEVAR, but the physiologic derangements appear to be less than with open surgery. Thus, there is a trade-off of risks, with evidence of SCI occurring in both groups.
It is an interesting observation that four of the five patients with SCI in our series presented with prior AAA repair and compromised internal iliac circulation in the context of extensive aortic coverage.27 In addition, most patients with observed deficits required coverage of the left subclavian artery, although in each case it had been perfused by a left carotid–subclavian bypass before implantation of the endograft in an effort to mitigate the effect of retrograde vertebral perfusion of the arm on SCI. Thus in our series, the left subclavian coverage was more likely a marker for extensive aneurysmal disease rather then as a source of steal for spinal cord perfusion.
The immediate management of patients presenting with SCI is critical. We used spinal drainage to 10 cm H2O in patients we considered to be at a relatively high risk for SCI (extensive coverage, prior AAA repair, or compromised internal iliac circulation). Should SCI symptoms manifest, the spinal drainage is increased by dropping the cutoff from 10 cm H2O to 5 cm or 0 until resolution of the neurologic symptoms is appreciated. Furthermore mean arterial pressures were maintained at >90 mm Hg, preferably without the use of vasoconstrictors that may detrimentally effect the collateral perfusion of the anterior spinal artery by constricting the small pelvic, lumbar, and intercostal vessels. Reversible deficits were observed in a few patients, and such maneuvers limited the symptoms to <24 hours; however, we are unsure whether such symptoms would have abated without these adjunct therapies.
Several reports have noted that both open and endovascular patients are at risk for stroke, yet the mechanism by which this complication occurs likely differs. Strokes after open surgery may be attributed to embolic events that occur with the placement of proximal cross-clamps but may also result from hypotension and hypoperfusion if circulatory arrest is used during the proximal reconstruction. Embolic CVAs have been linked with instrumentation of the aortic arch.4, 7
In two patients in our series who received regional anesthesia, emboli occurred after arch ballooning done in an effort to force the graft to conform to the lesser curvature. The results of ensuing cerebral angiography did not prompt neurointerventional rescue in either case. Subsequent computed tomography demonstrated multiple infarcts within the posterior circulation in both patients, implying a diffuse embolic process. Two of the three patients who survived to hospital discharge after TEVAR complicated by CVA ultimately died of strokes in the same cerebrovascular distribution. Did the arch ballooning that occurred in these two patients render plaques in the region unstable, prompting secondary CVAs? Was a region of turbulent flow created in the region of the proximal endograft or left subclavian artery? Should a more aggressive anticoagulation regimen be used in such circumstances? These questions will require further investigation, which will require pooling of patients with such complications owing to the relatively low incidence of this complication.
Why is the posterior cerebral vasculature so prominently affected? It seems logical that embolic debris would migrate to this territory given the proximity of the endograft placement to the vertebral artery. The use of cerebral protection during arch TEVAR has been suggested. However, given the location of strokes in our series, it would seem prudent to protect all four vessels feeding the intracranial circulation rather than the internal carotids only. This would add considerable complexity to the procedure and likely create problems related specifically to the embolic protection devices. Yet this complication remains devastating and, in our series, associated with a very high mortality.
Prompt treatment of type I and III endoleaks has been advocated to prevent rupture.7, 28 The management of type II endoleaks, particularly in the thoracic aorta, has not been fully elucidated; however, observation of these leaks is the rule rather then the exception.28
Migration has also been the subject of some controversy after TEVAR. There appears to be less debate about device migration in the abdominal aorta.29 This phenomenon is less well-defined in the thoracic aorta. The Talent (Medtronic/AVE, Santa Rosa, Calif) investigators30 defined migration as a ≥5 mm increased in the distance between the proximal covered part of the stent and left subclavian artery, both at implantation and during follow-up, yet they did not use 3D imaging studies to detect movement. The published TAG (W. L. Gore & Assoc. Flagstaff, Ariz) core laboratory data31 noted a 3-year freedom from migration of 83% but also relied on 2D studies with a variable z-plane resolution for analysis. We used much more rigorous methods of migration assessment, which we previously described.9
Although the use of active fixation mechanisms may help to discourage migration, the graft still requires the aorta to remain at a stable diameter within the sealing zones. We observed migration in four patients: two were related to distal migration of proximal stents, and two had proximal migrations of distal stents. It is interesting to evaluate the specifics of the migration cases. The distal stents that migrated involved implants that did not have distal active fixation mechanisms like those used in TX2D devices. One of the migrations occurred in the setting of a thoracoabdominal aneurysm where a polyester wrap was placed around an aneurysmal supraceliac aorta during the placement of an elephant trunk graft to “create” a landing zone, which ultimately proved to be unstable (Fig 4). Both patients with migration of the proximal stents, when the preoperative films were retrospectively analyzed with 3D tools, had unhealthy aortic regions where the initial sealing/fixation systems were deployed.
Life-table analysis of this patient series estimates that 93% of the patients will be free of migration at 5 years of follow-up. It seems likely that with more extended follow-up data, the incidence of migration will increase, adding importance to the detailed assessment of imaging studies at each follow-up visit. Such image analyses are also helpful to detect component separation. As the endovascular device conforms to the morphology of the aneurysm, the prosthesis generally lengthens to sit along the outer curvature of the aortic wall throughout the thoracic aorta. The increased prosthesis length must be derived either through proximal or distal fixation system migration or loss of overlap between components. Should two overlapped components separate such that a type III endoleak is created, repair should ensue. We observed this phenomenon in three patients, one of which had aortic growth that resolved after the addition of another overlap component.
Other device integrity problems were notably uncommon. No stent fractures or cases of fabric disruption occurred. Barb fractures were noted in seven patients, one of which was associated with migration. The cause and effect of the barb fracture and migration is unknown. The lack of aneurysmal growth after TEVAR with these devices in all but one patient, who developed a component separation that was treated and resolved, adds credence to the integrity of the fabric and overall repair.
A single-center study such as this has both advantages and limitations. Technical expertise and clinical judgment go hand in hand with volume and experience, and thus, these results are not representative of what may occur in a multi-institutional study with similar patients. It is likely that the patients treated in this series were anatomically more complex than patients in other reports simply owing to referral patterns.
A treatment bias exists, as in any nonrandomized study, with regard to the choice of available devices. Multiple studies occurred during patient enrollment whereby anatomically challenging patients were enrolled into this trial. However, with a single center and a well-developed research infrastructure, we were able to ensure that >90% of the patients had cross-sectional imaging data available at 2 years of follow-up, a level of patient accountability that to our knowledge has not been reported in any other series. In addition, all imaging studies were reviewed using a 3D workstation by at least three individuals blinded to each other's interpretations. Although our mean follow-up was just >3 years, with an upper limit of 6 years, these results still represent intermediate-term follow-up, and longer-term data are required.
Conclusions
The treatment of patients with thoracic pathology with the Zenith TX1 and TX2 devices is associated with reasonable intermediate-term results. The perioperative risks remain significant, and anatomic challenges still exist mandating the use of hybrid procedures in some circumstances. The incidence of late complications is relatively low. Late endoleak, component separation, and device integrity issues were uncommon, despite detailed radiographic evaluation in a study with excellent patient accountability. Migration was also rare and appeared to occur primarily when the fixation and sealing segments of the grafts were placed into relatively unhealthy aortas. Unquestionably, longer-term follow-up is necessary in addition to the accumulation of more patients with varying types of anatomy or disease etiology. Further work in this area will improve our understanding of thoracic aortic disease, judgment regarding patient selection, and choice of sealing and fixation zones.
Author contributions
References
- . Mid-term results for second-generation thoracic stent grafts. Br J Surg. 2003;90:811–817
- Endovascular stent-graft placement of aneurysms involving the descending aorta originating from chronic type B dissections. Ann Thorac Surg. 2007;83:1635–1639
- . Transluminal placement of endovascular stent-grafts for the treatment of descending thoracic aortic aneurysms. N Engl J Med. 1994;331:1729–1734
- Endovascular repair of thoracic aortic lesions with the Zenith TX1 and TX2 thoracic grafts: intermediate-term results. J Vasc Surg. 2005;41:589–596
- Nonsurgical reconstruction of thoracic aortic dissection by stent-graft placement. N Engl J Med. 1999;340:1539–1545
- . Endovascular treatment of thoracic aortic diseases: combined experience from the EUROSTAR and United Kingdom Thoracic Endograft registries. J Vasc Surg. 2004;40:670–679
- Endovascular treatment of thoracic aortic aneurysms: results of the phase II multicenter trial of the GORE TAG thoracic endoprosthesis. J Vasc Surg. 2005;41:1–9
- Endovascular aneurysm repair versus open repair in patients with abdominal aortic aneurysm (EVAR trial 1): randomised controlled trial. Lancet. 2005;365:2179–2186
- An evaluation of centerline of flow measurement techniques to assess migration after thoracic endovascular aneurysm repair. J Vasc Surg. 2006;43:1103–1110
- . Improved prognosis of thoracic aortic aneurysms: a population-based study. JAMA. 1998;280:1926–1929
- Endovascular stent graft repair for aneurysms on the descending thoracic aorta. Ann Thorac Surg. 1998;66:19–24
- Outcome and expansion rate of 57 thoracoabdominal aortic aneurysms managed nonoperatively. Am J Surg. 1995;170:213–217
- Prospective study of the natural history of thoracic aortic aneurysms. Ann Thorac Surg. 1997;63:1533–1545
- . Resection of the thoracic aorta with replacement by homograft for aneurysms and constrictive lesions. J Thorac Surg. 1955;29:66–100
- . Graft replacement of aneurysm in descending thoracic aorta: results without bypass or shunting. Surgery. 1981;89:73–85
- . Thoracoabdominal aneurysm repair: results with 337 operations performed over a 15-year interval. Ann Surg. 2002;236:471–479
- . Left heart bypass during descending thoracic aortic aneurysm repair does not reduce the incidence of paraplegia. Ann Thorac Surg. 2004;77:1298–1303
- Descending thoracic aortic aneurysm repair: 12-year experience using distal aortic perfusion and cerebrospinal fluid drainage. Ann Thorac Surg. 2005;80:1290–1296
- Evolution of risk for neurologic deficit after descending and thoracoabdominal aortic repair. Ann Thorac Surg. 2005;80:2173–2179
- Stent-graft placement in atherosclerotic descending thoracic aortic aneurysms: midterm results. J Endovasc Ther. 2004;11:26–32
- . Endovascular stent-graft management of thoracic aortic diseases. Eur J Radiol. 2001;39:42–49
- Endovascular stent graft placement in patients with acute thoracic aortic syndromes. Eur J Cardiothorac Surg. 2003;23:788–793
- Endovascular repair of lesions involving the descending thoracic aorta. J Vasc Surg. 2005;42:1063–1074
- . Endovascular stent grafting versus open surgical repair of descending thoracic aortic aneurysms in low-risk patients: a multicenter comparative trial. J Thorac Cardiovasc Surg. 2007;133:369–377
- Hybrid approaches to thoracic aortic aneurysms: the role of endovascular elephant trunk completion. Circulation. 2005;112:2619–2626
- Neurological complications following endoluminal repair of thoracic aortic disease. Cardiovasc Intervent Radiol. 2007;30:833–839
- Endovascular repair of descending thoracic aortic aneurysms: an early experience with intermediate-term follow-up. J Vasc Surg. 2000;31:147–156
- Endoleaks after endovascular repair of thoracic aortic aneurysms. J Vasc Surg. 2006;44:447–452
- Stent-graft migration: a reappraisal of analysis methods and proposed revised definition. J Endovasc Ther. 2004;11:353–363
- Results of endovascular repair of the thoracic aorta with the Talent Thoracic stent graft: the Talent Thoracic Retrospective Registry. J Thorac Cardiovasc Surg. 2006;132:332–339
- . Aortic neck morphology after endovascular repair of descending thoracic aortic aneurysms. J Vasc Surg. 2006;43:26–31
Competition of interest: Dr Greenberg receives grants and research support from Cook Inc, W. L. Gore & Assoc, and Terarecon, is a consultant for Cook Inc and Boston Scientific, and receives other financial or material support from Cook Inc, Intellectual Property. Dr Clair is a paid consultant for Cordis, an unpaid consultant for Timna and Minnow Medical, is a speaker for FoxHollow, Cook Inc, W. L. Gore & Assoc, and OmniSonics, and is an advisory board member for Medtronic and Boston Scientific. Dr Lyden is a consultant for Cook Inc.
PII: S0741-5214(08)00279-6
doi:10.1016/j.jvs.2008.02.028
© 2008 The Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.
Refers to erratum:
- Correction
