Endovascular treatment of thoracic aortic aneurysms: Results of the phase II multicenter trial of the GORE TAG thoracic endoprosthesis
Article Outline
- Abstract
- Methods
- Results
- Discussion
- Conclusion
- Appendix: Trial sites and principal investigators
- Discussion
- References
- Copyright
Objective
A decade after the first report of descending thoracic aortic aneurysm (DTA) repair with endografts, a commercial device is yet to be approved in the United States. The GORE TAG endoprosthesis, an investigational nitinol-supported expanded polytetrafluoroethylene tube graft with diameters of 26 to 40 mm, is the first DTA device to enter phase II trials in the United States and has been used worldwide for a host of thoracic pathologies.
Methods
A multicenter prospective nonrandomized phase II study of the GORE TAG endoprosthesis was conducted at 17 sites. Enrollment was from September 1999 to May 2001. Preoperative workup included arteriography and spiral computed tomography scans of the chest, abdomen, and pelvis. Follow-up radiographs and computed tomography scans were obtained at 1, 6, and 12 months and yearly thereafter.
Results
A total of 139 (98%) of 142 patients had a successful implantation of the device. Inadequate arterial access was responsible for the 3 failures. The mean DTA size was 64.1 ± 15.4 mm. Men slightly outnumbered women (57.7%), with an average age of 71 years, and 88% of the patients were white. Ninety percent were American Society of Anesthesiologists category III or IV. One device was used in 44% of patients, and 56% required two or more devices to bridge the thoracic aorta. The left subclavian artery was covered in 28 patients, with planned carotid-subclavian transposition. The procedure time averaged 150 minutes, estimated blood loss averaged 506 mL, intensive care unit stay averaged 2.6 days, and hospital stay averaged 7.6 days. Within 30 days, 45 (32%) patients had at least 1 major adverse event: 5 (4%) experienced a stroke, 4 (3%) demonstrated temporary or permanent paraplegia, 20 (14%) experienced vascular trauma or thrombosis, and 2 (1.5%) died. Mean follow-up was 24.0 months. Four patients had aneurysm-related deaths. Three patients underwent endovascular revisions for endoleak. No ruptures have been reported. Twenty wire fractures have been identified in 19 patients; 18 (90%) of these occurred in the longitudinal spine, and only 1 patient required treatment. At 2 years, aneurysm-related and overall survival rates are 97% and 75%, respectively.
Conclusions
The GORE TAG thoracic endoprosthesis provides a safe alternative for the treatment of DTAs, with low mortality, relatively low morbidity, and excellent 2-year freedom from aneurysm-related death. Longitudinal spine fractures have so far been associated with rare clinical events.
Descending thoracic aortic aneurysms (DTAs) are less prevalent than their abdominal counterparts and are esti-mated to affect 10 of every 100,000 elderly adults.1 Thirty to forty percent of these aneurysms occur exclusively in the descending thoracic aorta.1, 2 Open repair of DTAs has traditionally been a procedure associated with high morbidity and mortality and has been offered only to good surgical candidates. Perioperative death rates exceeding 10%, a risk of paraplegia of 4% to 5%, and a long recovery from thoracotomy make this a weighty undertaking for surgeon and patient alike.3, 4
Endovascular repair of thoracic aneurysms offers the benefit of aneurysm exclusion without the physiologic insult of prolonged proximal aortic clamping and direct surgical exposure, which may improve early morbidity.5, 6 Higher-risk patients previously denied open treatment might also benefit from this less invasive approach.
Although endovascular treatment of aneurysm disease was evaluated early in both the abdominal and thoracic aorta,7, 8 commercial development of thoracic endografts has lagged behind the rush to introduce stent grafts suitable for the more common abdominal aortic aneurysms. No thoracic devices have yet gained regulatory approval in the United States, whereas four abdominal products have been marketed since 1999. The GORE TAG thoracic endoprosthesis (TAG; W.L. Gore, Flagstaff, Ariz) was the first thoracic device to enter clinical trials in the United States, in 1998. The trial was halted in November 2001 after the discovery of longitudinal spine fractures. A new trial was initiated in late 2003 with a slightly modified device. This study reports the clinical results of the initial phase II multicenter trial of the original TAG device in the treatment of descending DTA.
Methods
Enrollment
Enrollment in this prospective, nonrandomized study occurred from September 1999 to May 2001 at 17 institutions across the United States (Appendix). Surgical candidates with a DTA twice as large as the normal thoracic aorta adjacent to the aneurysm underwent detailed preoperative spiral computed tomography (CT) scanning and diagnostic arteriography. Proximal and distal landing zones of a suitable diameter and at least 2 cm long were required. The proximal landing zone had to be distal to the left common carotid artery. If the left subclavian artery was covered to achieve this zone, then a carotid to subclavian artery bypass was performed. A 7% to 18% oversizing of endoprosthesis to artery was recommended by the device manufacturer. Select anatomic and medical inclusion and exclusion criteria are listed in Table I. A nonrandomized group of open-repair DTA patients from the same institutions were enrolled in the study and followed up. The surgical results of this group are not reported here because of regulatory requirements pending review by the Food and Drug Administration.
Table I. Anatomic and medical inclusion and exclusion criteria for the GORE TAG trial
| Inclusion criteria |
 Anatomic |
| Descending TAA at least twice the size of the normal thoracic aorta (fusiform aneurysm) |
| Minimum 2 cm of normal thoracic aorta proximal and distal to aneurysm |
| Aortic taper of no more than 4 mm, or ability to treat with more than one graft |
| Medical |
| Life expectancy >2 y |
| Surgical candidate |
| Male or infertile female >21 y |
| Able to comply with consent and follow-up |
| Exclusion criteria |
| Anatomic |
| Significant thrombus in landing zones |
| Mycotic aneurysm |
| Unstable patient with rupture |
| Acute or chronic dissection |
| Planned occlusion of carotid or celiac arteries to seal |
| Medical |
| Major operation (other than planned subclavian to carotid transposition or bypass) within 30 d |
| MI or CVA within 6 wk |
| Creatinine >2.0 mg/dL not on dialysis |
| Connective tissue disorder |
| Documented drug abuse within 6 mo |
| Participation in another investigational device or drug study within 1 y |
Follow-up protocol
CT scans, plain radiographs, and physical examinations were obtained at 1, 6, and 12 months and yearly thereafter. A 3-month visit with a CT scan was also included in patients with an early endoleak seen at 1 month. Clinical data were reported by individual centers and monitored by sponsor representatives. Major adverse events were adjudicated by a Clinical Events Committee, and source documents were reviewed by the authors for this report. A core laboratory separately reviewed all imaging studies. Major adverse events were defined as clinical events that required therapy or those that resulted in an unplanned increase in the level of care, prolonged hospitalization, permanent adverse sequelae, or death.9 Migration was defined as displacement of all or part of the device sufficient to be associated with another complication (eg, endoleak) or more than 1 cm of movement of all or part of the device as seen by CT or chest radiograph. Aneurysm sac size change was defined by a diameter increase or decrease of 5 mm or more. All imaging-related data are reported as collected by the core laboratory unless otherwise indicated. Clinical data reported here are from the compilation of monitored research forms from the sites except for changes made by the Clinical Events Committee in conjunction with the authors after careful review of source documents involved. Results are reported as mean ± SD (range) where appropriate. Kaplan-Meier curves and log-rank tests were used to plot survival over time, and the χ2 test was used to compare nominal data. P ≤ .05 was considered significant.
Device description and deployment
The TAG endoprosthesis is a flexible nitinol-supported polytetrafluoroethylene (PTFE) graft available in diameters of 26 to 40 mm and in 10-, 15-, and 20-cm lengths. During the trial, 7.5- and 12.5-cm lengths were also available but have since been discontinued. The exoskeleton is bonded to the graft material without sutures and is constrained by a PTFE sleeve. The device profile depends on the size of the graft and requires a 20F to 24F sheath. Deployment is very rapid and occurs with the release of the constraining sleeve. The TAG expansion is designed to start from the middle of the endoprosthesis and progress toward the ends of the graft (Fig 1). This is designed to avoid the displacing forces of the high arterial flow when the device is partially deployed in a standard deployment mode. The device is then additionally expanded with a specially designed trilobed balloon that allows flow to continue during inflation (Fig 2). If more than one device was used, the most proximal device was deployed first, and then the second was deployed distal to this with at least 2 cm of overlap. The original device used in the phase II trial reported here contained two longitudinal spines for columnar support. The device has been redesigned recently: the longitudinal spine was removed, and the expanded PTFE material was reinforced (Fig 3). All procedures were performed under fluoroscopy by using standard endovascular techniques.
Fig 1.
Deployment sequence of the GORE TAG device.
Fig 2.
Trilobed balloon with arrows illustrating flow-through channels.
Fig 3.
Original (left) and re-designed (right) GORE TAG devices.
The protocol and procedures of this trial were approved by the Food and Drug Administration and by the individual institutional review boards of each participating center. Patients signed a research informed consent form before participation in the study.
Results
Demographics
One hundred forty-two patients satisfied all inclusion and exclusion criteria and were enrolled and scheduled to receive the TAG thoracic endoprosthesis. The average aneurysm size was 64.1 ± 15.4 mm (range, 20-110 mm). Men slightly outnumbered women (58% vs 42%), with an average age of 71 years (range, 30-86 years), and 88% of the patients were white. Medical comorbidities are listed in Table II. One hundred seven patients (90%) were American Society of Anesthesiologists category III or IV.
Table II. Medical comorbidities of GORE TAG patients
| Parameter | No. of Patients (%) |
|---|---|
| Coronary artery disease | 70(49) |
| Cardiac arrhythmia | 34(24) |
| Congestive heart failure | 13(9) |
| Stroke | 14(10) |
| Symptomatic aneurysm | 32(23) |
| Aneurysm of traumatic origin | 8(6) |
| COPD | 56(39) |
| Smoking history | 118(83) |
| Dialysis | 3(2) |
| Paraplegia | 1(1) |
Procedure and hospital course
One hundred thirty-nine of 142 patients had successful deployment of their TAG device(s). The three failures of deployment were all secondary to access issues related to iliac vessel size and tortuosity. Twenty-one patients (15%) had a conduit placed to advance the device to the thoracic aorta. The location of the conduit varied from the external iliac artery to the infrarenal aorta to bypass inadequate access vessels from the femoral region. Two hundred thirty-seven devices were used to bridge the thoracic aneurysm: 62 (44%) patients received 1 device, 61 (43%) patients received 2 devices, 11 (8%) patients received 3 devices, and 5 (4%) patients required 4 devices. Procedure time averaged 150 ± 74 minutes (range, 61-580 minutes), with 506 ± 945 mL (range, 0-8000 mL) of blood loss. Time in the intensive care unit averaged 2.6 ± 14.6 days (range, 0-167 days), with a median of 1 day. The total hospital stay was 7.6 ± 18 days (range, 1-190 days), with a median of 3 days. The left subclavian artery was covered in 28 cases after a planned left carotid-subclavian bypass to preserve subclavian flow when the artery was covered by the thoracic stent graft. Sixteen carotid-subclavian bypasses were performed as staged procedures, and 12 patients had this operation completed on the same day as the thoracic stent placement. Two inadvertent side branch coverages were noted: one of the subclavian artery (requiring no treatment) and another of the abdominal visceral vessels. The latter patient underwent an open abdominal extraction of the misplaced device and redeployment of new devices across the thoracic aneurysm without trauma to the visceral vessels.
Adverse events
Early and late adverse events are detailed in Table III. Major adverse events within 30 days were noted in 45 (32%) patients, and several had more than 1. The most common were bleeding, cardiopulmonary events, and vascular trauma during the procedure. All vascular trauma events were in the perioperative period.
Table III. Early and late major adverse events
| Complications <30 d | No. of Patients (%) |
|---|---|
| Any major | 45(32) |
| Bleeding complications—all | 12(9) |
| Any endoleak | 5(4) |
| Pulmonary | 14(10) |
| Cardiac | 4(3) |
| Vascular trauma/thrombosis | 20(14) |
| Stroke | 5(4) |
| Paraplegia/paraparesis | 4(3) |
| Death | 2(1.5) |
| Late events | 1-y patients (%) | 2-y patients (%) |
|---|---|---|
| Migration—proximal* | 0/97(0) | 3/68(4) |
| Migration—components* | 1/84(1) | 4/61(6) |
| Endoleak rate* | 7/97(7) | 6/68(9) |
| Size decrease ≥5 mm* | 23/83(28) | 24/64(38) |
| Size increase ≥5 mm* | 6/83(7) | 11/64(17) |
| Endovascular revision | 1 | 1 |
| Conversion to open | 1(1) | 0 |
| Ruptures | 0 | 0 |
| Aneurysm-related death | 2(1.5) | 0 |
* Core laboratory reports. |
Aneurysm-related mortality
There were three in-hospital deaths, for an operative mortality of 2%. One early death was due to a postoperative stroke and another to a cardiac event that occurred on postoperative day 11. The third death occurred nearly 7 months after the procedure from a long complicated course following an anoxic brain injury after respiratory arrest. This patient eventually died of septic complications and was found at autopsy to have had an aortoesophageal fistula. The only other aneurysm-related death occurred 2 months after the original procedure after surgical conversion for aneurysm enlargement and suspected infection. At conversion, the patient was found to have an aortoesophageal fistula that was successfully excluded with graft removal and extra-anatomic bypass. On postoperative day 13, the patient experienced respiratory arrest and sustained an anoxic brain injury. He died 3 days later.
Cerebrovascular accidents
Five patients (3.5%) had a perioperative stroke. One stroke was fatal, and four patients were discharged to a rehabilitation facility or home. Three of five strokes affected both the anterior and posterior circulation distributions, and three were right-sided. Of the two left-sided strokes, one patient had a carotid-subclavian bypass. The strokes clustered in patients with proximal aneurysmal disease who were undergoing a carotid-subclavian bypass in preparation for the endograft placement. Four (80%) of the five patients with a stroke had a planned bypass either as a staged intervention (n = 2) or in conjunction with the endovascular procedure (n = 2). In total, 4 of 28 patients undergoing the carotid-subclavian intervention (14%) had a cerebrovascular event after the TAG procedure, compared with 1% of patients who did not have a carotid-subclavian bypass (P < .001).
Spinal ischemia
Spinal drainage was not routinely used during the interventions. Four patients had spinal cord events in their postoperative course, and three had spinal drains placed at the time the deficits were noted. Immediately after the procedure, one patient was noted to have bilateral lower extremity weakness. Despite spinal drainage, the deficits persisted, and she was discharged in this condition. A second patient exhibited paraplegia 6 hours after surgery in conjunction with a hypotensive episode. These symptoms improved after the patient was stabilized and a drain was placed. At last follow-up, this patient was ambulating with a walker. A third patient exhibited lower extremity weakness and numbness on postoperative day 1. She improved by the time of discharge and was near baseline at 1 month after her procedure. A fourth patient was noted on postoperative day 1 to have decreased strength in the proximal muscles of the legs. This was treated with aggressive blood pressure support and had improved 24 hours later. She was discharged to a rehabilitation facility with good strength in all extremities. It is of note that three of four patients had more than one device implanted (two, three, and four devices). Two of the four patients also had an abdominal aneurysm previously treated with aortobifemoral bypass. Forty-two (29.6%) TAG patients had previous infrarenal aortic replacement, resulting in a 4.7% rate of spinal cord events in this subgroup. Paraplegia in patients without abdominal aortic replacement was 2% (P = .36; not significant).
Endoleaks
Of five early endoleaks at 30 days, one patient had a proximal type I endoleak and was treated with endovascular revision and additional grafts. This patient had an additional endovascular revision after 24 months of follow-up. The remaining endoleaks were thought to be type II. At 1 and 2 years, the rates of observed endoleaks were 7% and 9%, respectively. Two more patients were treated by endovascular revision: one during the second year and one during the third year of follow-up. Because endoleaks were not consistently observed in the same patient at all times, 21 patients (15%) had an endoleak at some time during the first 2 years of follow-up. Most endoleaks were of undetermined origin on CT scan. Three patients who had endoleaks at some time were also found to have sac expansion of 5 mm or more.
Follow-up
Average follow-up for this report was 728 days (24.0 months), with a range of 3 to 1626 days. Fig 4 depicts a Kaplan-Meier survival curve showing freedom from all mortality and aneurysm-related mortality. At 2 years, the rates are 75% and 97%, respectively. No ruptures were reported.
Fig 4.
Kaplan-Meier curves of aneurysm-related and non–aneurysm-related survival.
Late events are also noted in Table III. Four endovascular revisions in three patients and one surgical conversion were reported during follow-up; two of the revisions occurred after 24 months of follow-up. At 2 years, three proximal and four component migrations were noted by the core laboratory, without associated clinical events. Aneurysm sacs decreased in size in 24 (38%) of 64 patients, and size increased in 11 (17%). Three of the 11 patients with sac expansion had endoleaks at some point during follow-up. Twenty fractures have so far been identified in 19 patients at 2 years. Eighteen (90%) of 20 fractures were in the longitudinal spine, and only 2 were in the apical nitinol support rings. One patient had a fracture in the spine and the ring. Only one patient with a fracture has so far developed a clinical event. This was a type 3 endoleak corrected by an endovascular deployment of a new device.
Discussion
The natural history of DTAs and the morbidity and mortality of open repair are not well known, because most series combine them with more extensive thoracoabdominal aneurysms. Perioperative mortality as high as 10% and paraplegia rates approaching 4% to 5% have been reported in selected series.3, 4 These risks can be prohibitive in an elderly population. However, with a greater than 30% 5-year risk of rupture or death when the aneurysm is larger than 6 cm, there is a clear need for prophylactic repair of large aneurysms.10 Endovascular treatment of DTAs may offer both short- and long-term advantages when compared with open surgery, because it avoids a thoracotomy and thoracic aortic clamping.
The feasibility of endovascular treatment of DTAs was first reported in a high-risk population by Dake et al7 in 1994, who used homemade devices consisting of Dacron-covered (DuPont, Wilmington, Del) Gianturco Z-stents. As experience accumulated, significant limitations of a stiff device with a large introducer size became evident.11 Long-term results in high-risk patients, especially those deemed inoperable or those requiring arch deployment after a carotid subclavian bypass, were far less than acceptable, which led the group to nearly abandon this first-generation device.5 Other groups using similar constructions also encountered serious complications, including a mortality rate as high as 25%, late ruptures, and migrations.12
The advent of commercially manufactured grafts based on technology developed for the treatment of abdominal aneurysms brought the promise of a smaller, more flexible, and more durable prostheses. Experience with these devices is accumulating worldwide in a variety of clinical settings. Indications have been expanded to include dissections, traumatic injuries, and other thoracic pathology. Early single-center results are promising, with improved mortality and morbidity.6, 13 Criado et al,6 using the Talent device (Medtronic AVE, Santa Rosa, Calif), and White et al,14 using the AneuRx device (Medtronic AVE), both reported a 4% aneurysm-related mortality in series of both DTAs and dissections. Only 1 patient of the 74 treated by these 2 groups developed paralysis.
The TAG device was the first commercial device to enter trials in the United States. Single-center experiences in high-risk applications have confirmed a low mortality of 4% to 6% and a very low paraplegia rate.15, 16 The phase II results of the TAG device presented here illustrate a safe and technically successful means of treating a larger population of selected patients with DTAs. These elderly patients with significant cormorbidities, who were overwhelmingly American Society of Anesthesiologists class III or IV, had very low mortality and spinal ischemia rates of 2% each while having a high rate of successful deployment and aneurysm exclusion. These encouraging early and mid-term results cannot at present be compared with the “surgical control group” because of regulatory requirements, but they seem to improve on historical standards.
The TAG device has many technical advantages, including a very rapid deployment from the center of the graft out, which has eliminated the need for temporary cardiac arrest with adenosine.17 In addition, the flexibility of the TAG catheter is well suited to navigate tortuous anatomy and to be deployed in the aortic arch. The proximal end of the device was deployed in this series at the level of the left carotid artery in 20% of patients. This manipulation in the arch, however, was probably responsible for a significant incidence (nearly 14%) of strokes in this subgroup of patients. The etiology was most likely embolic, because the cerebral events were frequently multicentric and involved the posterior circulation in at least three of five patients. The associated carotid subclavian bypasses performed as part of this protocol were probably only a marker for the site of deployment and the associated risk, rather than the cause of the strokes. Whether such an adjunctive procedure is necessary in all cases is open to debate, because many European centers have experienced few complications from simply covering the left subclavian artery during such procedures.18
Although spinal cord events are always tragic, their 3% rate noted with this phase II trial is quite low and is less than many other reported ranges.12, 15, 19 Three of four patients with lower extremity weakness had significant improvement within short-term follow-up, and one had permanent deficits. Two of the patients had previous abdominal aneurysm surgery, and three had more than one graft placed to cover a significant portion of the descending thoracic aorta. These observations mirror other results showing that patients with previous aortic surgery and with long segments of covered thoracic aorta have an increased incidence of spinal cord events after thoracic endografting.19, 20
Stent fractures, which led to the interruption of this development program for more than 2 years, have proved to be less of a clinical concern than initially anticipated. Twenty fractures in 19 patients have been reported thus far in patients enrolled in this phase II trial, and only 1 patient has had a clinical event related to the fracture. Metal fatigue and material deterioration have been reported with virtually every endograft used to date21 and are known complications of endoluminal treatment and product evolution. Stent fractures seen with the TAG have been overwhelmingly in the spine of the graft (18 of 20)—a feature that has now been eliminated in the modified device in current trials. Additional expanded PTFE thickness and reinforced binding to the metal stent are expected to compensate for the support previously provided by the longitudinal spine and to thus avoid the possible side effects of fracture. This revised design is currently under investigation. Because most patients with spine fractures were asymptomatic, the expectation is that the redesigned product will perform similarly to the original product, at least in the early follow-up phase. A finding that warrants close attention is the number of patients who had sac enlargement after successful DTA endograft placement. Seventeen percent of patients showed 5 mm or more of sac enlargement at the 24-month follow-up. The significance of this finding is unknown. These are patients who have not all demonstrated endoleak yet have expanding sac size. This phenomenon has been reported with the Excluder (W.L. Gore) abdominal endoprosthesis22 and as yet has had no dire clinical effects. The new modified TAG device has reduced porosity, which may be of benefit for the cases of sac enlargement.
Access to the thoracic aorta remains a significant issue with endoluminal repair and accounts for all technical failures in this report. Twenty patients had vascular complications even though 15% of patients required a conduit to a larger part of the arterial tree. These rates are similar to those in other reported series6, 12, 13 and underline the importance of using conduits as a preventive measure rather than a bailout after arterial injury has been noted. This maneuver is quite important because sex distribution is far more even with thoracic aneurysms compared with their abdominal counterparts, as evidenced by 43% of the patients in this trial being women.
Future directions for thoracic stent grafting with the TAG device may include use in rupture, acute type B dissection, and a host of other thoracic pathologies. As experience and comfort with the device increase, there may be wider applications that could decrease the morbidity associated with open operations.
Conclusion
The TAG thoracic endoprosthesis provides a safe alternative for the treatment of DTAs, with low mortality, relatively low morbidity, and an excellent 2-year freedom from aneurysm-related death. Longitudinal spine fractures have so far been associated with rare clinical events that will, it is hoped, be eliminated by the modified version presently undergoing clinical testing.
Appendix: Trial sites and principal investigators
| Institution | City | Investigator(s) |
|---|---|---|
| Baylor | Houston, Tex | Joseph Coselli, MD |
| Cleveland Clinic | Cleveland, Ohio | Roy Greenberg, MD |
| Emory | Atlanta, Ga | Elliot Chaikof, MD |
| Hartford Hospital | Hartford, Conn | Micheal Hallisey, MD |
| Johns Hopkins | Baltimore, Md | Lawrence Hoffman, MD |
| Massachusetts General | Boston, Mass | Richard Cambria, MD |
| Mayo Clinic | Rochester, Minn | Kenton Zehr, MD |
| Mt Sinai | New York, NY | Larry Hollier, MD |
| Northwestern | Evanston, Ill | Mark Marasch, MD |
| Stanford | Stanford, Calif | Steven Kee, MD |
| R. Scott Mitchell, MD | ||
| University of Florida | Gainesville, Fla | James Caridi, MD |
| James Seeger, MD | ||
| University of Michigan | Ann Arbor, Mich | David Williams, MD |
| University of Pennsylvania | Philadelphia, Pa | Joseph Bavaria, MD |
| Jeffrey Carpenter, MD | ||
| University of Pittsburgh | Pittsburgh, Pa | Michel Makaroun, MD |
| University of South Carolina | Charleston, SC | Renan Uflatker, MD |
| John Ikonomedis, MD | ||
| Washington University | St Louis, Mo | Gregorio Sicard, MD |
| Yale | New Haven, Conn | John Elefteriades, MD |
Discussion
Dr R. Greenberg (Cleveland, Ohio). As the invited discussant, I think it is imperative that I disclose that I am an investigator in this trial and the other two commercial trials and that I have a development agreement with Cook, Incorporated. Assuming that this study is designed as other abdominal aortic aneurysm predecessor studies have been designed, intentionally to avoid a rupture-free survival end point, we've placed a heavy reliance on surrogate end points, and we must carefully consider these statistical analyses. In the study design inherent with this article, the morbidity data collected in a nonrandomized fashion and compared with historical controls must be called into question. There are no concurrent controls, and there are potentially anatomic differences between these groups. I would like you to address that. Assuming that we agree that these comparisons are valid, we then need to turn our attention to the surrogate end points. I would ask you what surrogate end points you think are most predictive of long-term rupture-free survival and the lack of need for secondary interventions?
Our own conclusions have come to the assessment of both sac size and migration data. In your article and presentation, these circumstances were analyzed. However, the circumstances were really analyzed by using, at least from what I can gather, two-dimensional assessments of these criteria. In our experience, the centerline of flow and three-dimensional offline reconstruction have been imperative in the analysis of thoracic data.
The sac growth is also concerning and perhaps parallels the sister abdominal device noted at 7% and 17% at 2 years. I question whether this is different from the natural history of such aneurysms without treatment and whether this is going to create long-term complications.
Last, I would like you to comment on the two aortoesophageal fistulas that developed in the first year after implantation, as reported in the article, and perhaps you can elaborate on the etiology of the strokes—whether they were thrombotic, embolic, or otherwise created.
Dr Makaroun. The first question regards the comparison with controls. As you can tell from my presentation, I strictly avoided any comparison with controls, whether they were historic or otherwise.
As far as the surrogate markers for the clinical success of the therapy, including the size and the migration, the evaluation of these by the core laboratory, as you do mention, is fraught with a lot of sources of error. And, if you note, the numbers have changed from the abstract, the numbers at 1 year until this presentation, after additional detailed studies were re-evaluated. The size increase that is noted at 1 and 2 years is slightly higher than you may want to expect in a trial of this sort, but I have no way of telling whether this is close to the natural history of these particular patients. It is slightly lower than the abdominal component at this time frame, but hopefully the change in the fabric that was instituted in the new device will have different results in the future.
As far as the two aortoenteric fistulas that were described, the first one is the one that led to the conversion and the death at 3 months, and the second one was actually discovered only at autopsy in the patient who was in the hospital for 6 months and died from multiple septic problems.
Dr K. Cherry (Charlottesville, Va). I am intrigued about the strokes. Presumably, the carotid subclavian bypasses were done to prevent ischemia to the left subclavian, yet they seemed to be incriminated a bit in the stroke rate. And it's a follow-up question. When were they done temporally in relation to the thoracic reconstruction? And when was the stroke in relation to either the bypass or the construction? Did you feel that the carotid subclavian bypass played a role in the stroke?
Dr Makaroun. The five strokes that we identified were associated with carotid subclavian bypasses, but we are not implicating the bypass in the causation of the stroke but as an indicator of the fact that it was required to deploy the device in the thoracic arch. Unfortunately, despite having reviewed the source documents for these, it is very hard to decide whether these were truly embolic or thrombotic in nature.
The carotid subclavian bypasses were not all performed in conjunction with the deployment of the device. Of the four strokes, one was actually performed as a staged procedure before the deployment of the device, and three of the four actually had the carotid subclavian bypass or the transposition performed at the same time. I was trying to point out only that the carotid subclavian bypass is an indicator that when you are manipulating the arch with both the wires and the deployment, you carry a significantly higher chance of provoking a stroke than when the deployment is strictly in the descending aorta past the subclavian.
Dr J. Parodi (St Louis, Mo). I have a short comment and a question. The comment is that I believe that surgeons should be prepared for rescue procedures and not accept the consequences of stroke without attempting this. We have done this in eight patients. And the question is, in those cases of stroke, have you attempted to study those patients and attempted any rescue procedure?
Dr Makaroun. I don't think that in any of the five patients a rescue procedure was performed. And I failed again to answer one of the questions of Dr Cherry, about whether the timing of the strokes was not after the staged carotid subclavian bypasses. Typically, they happened after the deployment of the device when the patient woke up from anesthesia. Most of these patients were not monitored to recognize the stroke intraoperatively. And a lot of them were performed under general anesthesia. So it was not recognized until the patient was waking up and the deficit was recognized, and none had an attempted rescue.
References
- . Improved prognosis of thoracic aortic aneurysms: a population-based study . JAMA . 1998;280:1926–1929
- . Thoracic aortic aneurysms: a population based study . Surgery . 1982;92:1103–1108
- . Surgical correction of descending thoracic aortic aneurysms under simple aortic cross-clamping . J Vasc Surg . 1989;9:568–573
- . Thoracoabdominal and descending thoracic aortic aneurysm surgery in patients aged 79 years or older . J Vasc Surg . 2002;36:469–475
- Midterm results of endovascular repair of descending thoracic aortic aneurysms with first-generation stent grafts . J Thorac Cardiovasc Surg . 2004;127:664–673
- . Stent graft repair in the aortic arch and descending thoracic aorta: a 4-year experience . J Vasc Surg . 2002;36:1121–1128
- . Transluminal placement of endovascular stent-grafts for the treatment of descending thoracic aortic aneurysms . N Engl J Med . 1994;331:1729–1734
- . Transfemoral intraluminal graft implantation for abdominal aortic aneurysms . Ann Vasc Surg . 1991;5:491–499
- . Reporting standards for clinical evaluation of new peripheral arterial revascularization devices . J Vasc Interv Radiol . 1997;8:137–149
- . Natural history of thoracic aortic aneurysms: indications for surgery, and surgical versus nonsurgical risks . Ann Thor Surg . 2002;74:S1877–S1880
- Endovascular stent-graft repair of thoracic aortic aneurysms . J Thorac Cardiovasc Surg . 1996;111:1054–1062
- Endovascular repair of descending thoracic aortic aneurysms: an early experience with intermediate-term follow-up . J Vasc Surg . 2000;31:147–156
- Evolving experience with thoracic aortic stent graft repair . J Vasc Surg . 2002;35:1–8
- Endovascular exclusion of descending thoracic aortic aneurysms and chronic dissections: initial clinical results with the AneuRx device . J Vasc Surg . 2001;33:927–934
- Challenges of endovascular tube graft repair of thoracic aortic aneurysm: midterm follow-up and lessons learned . J Vasc Surg . 2003;38:676–683
- . Endoluminal stent grafting of the thoracic aorta: initial experience with the Gore Excluder . J Vasc Surg . 2002;35:1163–1170
- . Adenosine-induced transient cardiac asystole enhances precise deployment of stent-grafts in the thoracic or abdominal aorta . J Endovasc Surg . 1996;3:270–272
- Initial experience with intentional stent-graft coverage of the subclavian artery during endovascular thoracic aortic repairs . J Endovasc Ther . 2002;9:SII39–SII43
- Risk of spinal cord ischemia after endograft repair of thoracic aortic aneurysms . J Vasc Surg . 2001;34:997–1003
- . Simultaneous abdominal aortic replacement and thoracic stent-graft placement for multilevel aortic disease . J Vasc Surg . 1997;25:332–340
- Mechanical failure of prosthetic human implants: a 10-year experience with aortic stent graft devices . J Vasc Surg . 2003;37:16–26
- . Late abdominal aortic aneurysm enlargement after endovascular repair with the Excluder device . J Vasc Surg . 2004;39:1236–1242
Competition of interest: Dr Makaroun received a grant support from W.L. Gore and has acted as a consultant for W.L. Gore. S.T.K., G.S., E.C., J.B., D.W., and R.P.C. received grant support from W.L. Gore. R.S.M. received grant support from W.L. Gore and has acted as a consultant for W. L. Gore.
PII: S0741-5214(04)01485-5
doi:10.1016/j.jvs.2004.10.046
© 2005 The Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.
