|
|
||||||||
Department of Cardiothoracic and Vascular Surgery, Deutsches Herzzentrum Berlin, Berlin, Germany
Accepted for publication April 13, 2009.
* Address correspondence to Dr Zipfel, Department of Cardiothoracic and Vascular Surgery, Deutsches Herzzentrum Berlin, Augustenburger Platz 1, Berlin, 13353, Germany (Email: zipfel{at}dhzb.de).
Presented at the Poster Session of the Forty-fifth Annual Meeting of The Society of Thoracic Surgeons, San Francisco, CA, Jan 26–28, 2009.
| Dr Zipfel discloses that he has a financial relationship with Jotec GmbH and Bolton Medical.
|
| Abstract |
|---|
|
|
|---|
Methods: In 102 patients treated with endovascular stent grafts, the LSA was covered. In a nonrandomized clinical practice, unprotected occlusion of the LSA was performed in 63 patients (61%), whereas 39 patients underwent extrathoracic subclavian to carotid artery revascularization before (n = 28) or concomitantly with (n = 11) the endovascular procedure.
Results: Left cerebral ischemia occurred in 11% of the unprotected group and in 5% of the protected group. The difference was not statistically significant. The difference in spinal cord ischemia was insignificant owing to the low incidence in general, but the covered length of the aorta was significantly longer in the protected group. Arm ischemia after unprotected LSA occlusion occurred in 25%.
Conclusions: The interpretation of the results remains speculative because many factors contribute to left cerebral ischemia. However, in terms of overall complications, there is a significant difference in favor of the group protected by revascularization of the LSA either before or simultaneously with stent grafting. Arm ischemia is mostly mild and can be managed secondarily. Subclavian revascularization is associated with relatively low risk and should be considered in advance, at least when extended covering of the thoracic aorta is intended.
For safe fixation of stent grafts in the distal aortic arch, it is often necessary to use the segment between the left common carotid artery (LCCA) and the left subclavian artery (LSA) as the proximal landing zone [1] with subsequent occlusion of the LSA. Liberal use of this technique is recommended in all cases with short distance between the LSA and the aortic lesion or in elongated aortic arches to avoid migration of the stent graft and subsequent type I endoleak, or in traumatic transsection to achieve safe exclusion of the contained rupture [2]. Based on the experience of sacrificing the LSA in coarctation repair [3, 4], LSA occlusion was initially deemed to carry a low risk, and clinical experience seemed to confirm that [5, 6]. When more experience was gained, some cases were noted with ischemic complications to which LSA occlusion may have contributed. Therefore, we changed our policy toward more liberal prophylactic LSA reconstruction. This retrospective study evaluates whether this prophylactic procedure may benefit the patient.
| Patients and Methods |
|---|
|
|
|---|
Data were gathered retrospectively from a Microsoft Access prospective database of all patients undergoing endovascular aneurysm or dissection repair. Stent grafts were only implanted when they were approved in the European Union. The prospective database was approved by the Institutional Ethics Committee. Written informed consent to data collection was obtained. This retrospective study was approved by the Institutional Ethics Committee, which waived the need for additional patient consent to the study.
Ninety-two procedures were performed in a standard operating room, and a surgical C-arm with angiography equipment (BV 300; Philips, Eindhoven, Netherlands) was used for intraoperative imaging. One procedure was performed in the cardiology angiography suite. Since March 2008, 9 procedures have been performed in a new hybrid operating room equipped with a fixed angiography unit with integrated angiography table (Artis Zee ceiling-mounted; Siemens, Erlangen, Germany). General anesthesia was used in all cases.
We used E-vita stent grafts (Jotec, Hechingen, Germany) in 52 procedures; Relay (Bolton Medical, Sunrise, FL) in 26; Talent (Medtronic Vascular, Santa Rosa, CA) in 23 procedures; and TAG (W. L. Gore Associates, Flagstaff, AZ) in 1 procedure. All stent grafts are self-expanding and are oversized by 10% to 20% in relation to the outer diameter of the aorta at the landing zone determined by preoperative computed tomography or magnetic resonance imaging. They are packed in delivery catheters of 22F to 27F. The stent grafts were advanced in retrograde manner from the femoral or iliac vessels and deployed as guided by the landmarks of a target angiogram. We have described the stent grafts and the implantation technique previously in detail [7].
Extrathoracic LSA revascularization was performed through a standard horizontal supraclavicular incision. The LCCA was exposed after mobilization of the internal jugular vein; then the fat tissue of the supraclavicular fossa was divided, and the phrenic nerve identified. To expose the LSA, the anterior scalenus muscle was partially or completely divided. When LSA to LCCA transposition was performed, the LSA was exposed further proximally to the origin of the left vertebral artery, as close to the aortic arch as possible. The LSA was then clamped and divided, and the stump was closed with a double 4-0 polypropylene mattress and running suture. The LSA was then moved to the LCCA and implanted with a rectangular end-to-side anastomosis with a 4-0 polypropylene running suture. In LCCA-LSA bypass procedures, this proximal dissection of the LSA was avoided, and an 8-mm collagen sealed polyester graft (FlowNit Bioseal; Jotec, Hechingen, Germany; or Hemashield Gold; Maquet, Rastatt, Germany) was implanted between the most cranial section of the LSA and the LCCA with two end-to-side anastomoses with 4-0 polypropylene running sutures, the LCCA anastomosis rectangular to the lateral aspect of the LCCA. Thirty bypasses were performed, with ligation of the proximal LSA in 6 cases and 9 transpositions. In simultaneous procedures, the LSA reconstruction was performed as the first step before stent graft implantation. The bypass or the transposed LSA was used as the approach for a separate catheter for target angiography (Figs 1 and 2)
[7, 8].
|
|
All ischemic neurologic events with a potential relationship to impairment of the LSA circulation are reported in this retrospective review, namely, any left cerebral reversible ischemia or stroke and any not clearly trauma-induced clinical spinal cord compromise.
| Results |
|---|
|
|
|---|
|
In nonrandomized clinical practice, unprotected occlusion of the LSA was performed in 63 patients (61%), whereas 39 patients underwent extrathoracic subclavian to carotid artery reconstruction before (n = 28) or concomitantly with (n = 11) the endovascular procedure (protected group).
Implant details of the two groups with protected or unprotected coverage of the LSA and ischemic complications to which the occlusion of the LSA may have contributed are given in Table 2. There was a 25% incidence of left arm ischemia in the unprotected group. A more detailed description of the neurologic events is given in Table 3. Overall incidence of left cerebral events was 9% (n = 9), with insignificant differences between the two groups (11% versus 5%). Incidence was 12% among trauma patients (n = 3 of 26), 11% (n = 4 of 37) among those with aneurysms and 5% (n = 2 of 39) among those with dissections. Two spinal cord events are reported although they were most likely related to the underlying disease or trauma. One case of spinal cord ischemia was induced by the acute dissection with malperfusion and was reversible within 24 hours after implantation of the stent graft with unprotected LSA occlusion (patient 9). In the second case (patient 10), the patient had posttraumatic reversible tetraplegia after stable cervical spine fracture without displacement, which was deemed spinal contusion by the orthopedic surgeons, but the contribution of LSA coverage by the stent graft remains unclear. Two cases of clearly trauma-related paraplegia in severely dislocated spinal fractures were excluded. One case was a patient (patient 5) who later had cerebellar and brainstem infarction and a progression of the spinal cord injury to the level C4/5. The other case was a 17-year-old boy, who was the only trauma patient with simultaneous LSA reconstruction, which was performed with the intent of preserving the arm circulation for pushing the wheelchair; interestingly, this patient did not experience any deterioration of his initial neurologic status. One multiple trauma patient with severe cerebral edema and final brain death after severe head injury was not considered as having an ischemic neurologic complication because it was obviously trauma induced.
|
|
|
| Comment |
|---|
|
|
|---|
Impairment of the Left Arm Circulation
Because of preformed collaterals, occlusion of the LSA orifice remains asymptomatic in many cases. Arm ischemia is mostly mild. Clinical problems have been reported to be infrequent in the literature. Many authors claim that if weakness or numbness of the arm occurs, that may disappear with time [6, 9]. In a recently published meta-analysis [10], left arm claudication was reported in 23 of 29 articles cited, with an incidence of 0% in 15 papers and an incidence between 2% and 14% in eight papers. We observed an incidence of 25% (n = 16) in the unprotected group. This obvious difference may be explained by a straightforward surgical approach to manage postoperative exercise-induced left arm weakness or pain after occlusion of the LSA. If the patients experience symptoms, we offer them secondary revascularization. Therefore, the majority were managed by secondary revascularization, with immediate resolution of the symptoms. Arm ischemia is less frequent in patients with thoracic aneurysms compared with dissection and trauma, probably owing to the higher age in this group that limits activity and thus exercise-induced symptoms.
Cerebral Ischemia
Many factors apart from LSA occlusion, such as extended atherosclerotic aortic arch disease, emergency procedures and trauma, may contribute to left cerebral ischemia. Stroke occurring after stent grafting may be caused as well by occlusion of the supra-aortic vessels or by emboli arising from manipulation of the aortic arch [11]. Therefore, in no case of this clinical study can LSA occlusion be proven as the single cause for cerebral ischemia. Three cases of left posterior cerebral infarctions were probably related to impairment of the perfusion of the vertebral artery. They were verified by cerebral computed tomography scans and differentiated from additional sequelae of head injury in 2 patients. In the remaining 6 cases (Table 3), the contribution to LSA occlusion is speculative. The two type-B dissection cases (patients 1 and 6) had no involvement of the arch vessels. A correlation to the underlying disease cannot be found, probably because of the small number and the various mechanisms. Also, the difference in the rate of cerebral ischemia with potential contribution to LSA occlusion of 11% of the patients with simple LSA coverage compared with 5% of patients with previous revascularization is insignificant, but ischemic complications in the protected group occurred in those patients with extensive coverage and revascularization of supra-aortic arteries. The stroke in 1 patient (patient 7) was caused by the revascularization procedure itself (plaque embolization from the LCCA anastomosis). Therefore, it is supposed that unprotected LSA occlusion by a stent graft may bear a higher risk of left cerebral ischemic complications. Peterson and colleagues [12] found similar results, with a 50% cerebral complication rate in 8 unprotected LSA occlusions in a study comparable to this one. Data from the open Eurostar registry of 606 patients revealed significantly different postprocedural stroke rates of 8% in the unprotected group and 0% in the protected group [13].
We did not see symptomatic subclavian steal syndrome with dizziness on exertion of the arm. Follow-up duplex sonography revealed reversal of flow in the vertebral artery in most cases [14], which seems to be normal after occlusion of the LSA, but symptoms requiring surgical intervention were not reported.
Diagnostic evaluation of the cerebral perfusion with extra- and intracranial Doppler sonography is recommended by several authors before occluding the LSA with a stent graft [9]. This evaluation is complex, time-consuming, and examiner-dependent. Therefore, when in doubt or without these diagnostic tools, protective revascularization is reasonable. Atherosclerotic disease of the common carotid artery at the site of the anastomosis should be ruled out by simple duplex ultrasound.
Spinal Cord Ischemia
The collateral perfusion of the spinal cord is fed by branches of the subclavian artery. In normal anatomy, the anterior and posterior spinal arteries arise from the vertebral arteries, but variations of direct origin of the spinal cord artery from the LSA have been described [15]. In these instances, the spinal cord perfusion is directly dependent on the LSA. Also, experimental findings support the theory that the LSA contributes to spinal cord perfusion. In a pig model, fewer intercostal arteries had to be occluded to induce paraplegia if in addition the subclavian arteries were occluded [16]. Experimental and clinical findings stress the importance of collateral blood flow to the spinal cord when intercostal arteries are sacrificed in conventional thoracoabdominal surgery [17, 18] and challenge the concept of reimplantation of intercostal arteries in conventional thoracic surgery [19]. Long stent grafts cause unavoidable extensive segmental artery sacrifice, and the incidence of spinal cord ischemia is unexpectedly low in our own and all other published series [8, 20–23]. Maintaining the collateral blood flow by avoiding cross-clamping and hypoperfusion in the endovascular technique is the most probable explanation. The first clinical evidence that LSA occlusion may be a risk for spinal cord ischemia in thoracic stent grafting was published by the Eurostar registry in 2007 [13]: in 606 patients, LSA occlusion was found to be an independent risk factor for spinal cord ischemia (odds ratio 3.9, p = 0.027).
Guided by these considerations, we practiced prophylactic revascularization at least in all patients in whom the placement of long stent grafts covering the entire descending thoracic aorta was planned and in all patients with previous replacement of the abdominal or thoracic aorta. Fortunately, we were not able to test the validity of this practice, because the rate of paraplegia is so low that subgroup analysis is statistically insignificant. Moreover, in both registered cases of spinal cord ischemia, it remains questionable whether the neurologic damage was caused by implantation of the stent graft.
Although, as a result of our policy, the coverage of the thoracic aorta was significantly longer in the protected group, with 23% coverage of the entire descending aorta from the LSA to the celiac artery, no spinal cord ischemia occurred. One remarkable case with additional abdominal aneurysm repair has been described earlier [24]. In conclusion, maintaining the collateral perfusion fed by the LSA seemed to prevent spinal cord ischemia.
Type II Endoleak
If the LSA origin is situated very close to the aortic aneurysm retrograde, endoleak [25] is anticipated. A typical situation is illustrated in Figure 4. The safest method to prevent that is extrathoracic ligation or disconnection of the LSA, which is then easily combined with revascularization [26]. Therefore, in these special cases, prevention of a type II endoleak is an additional indication for prophylactic LSA revascularization. Then, LSA transposition is the most reliable method because the proximal LSA is definitely disconnected. Bypass can be combined with ligation of the LSA proximal to the vertebral and mammarian artery. This procedure requires the same extent of surgical dissection as transposition. When retrograde leak is unlikely or probably minor, we prefer simple LCCA to LSA bypass. If an endoleak occurs, this can be closed with coils or occluders during the stent graft procedure.
|
Conclusion
Occlusion of the LSA with the stent graft may be safe if a short stent graft is implanted in the proximal aorta, and if stenoses and abnormalities of the supra-aortic and intracranial arteries supplying the brain can be ruled out. Arm ischemia can be handled safely after the procedure. We recommend prophylactic revascularization for all patients in whom the placement of long stent grafts covering the entire descending thoracic aorta is planned and for all patients with previous replacement of the abdominal or thoracic aorta. Protective LCCA to LSA bypass is mandatory in patients with previous coronary bypass grafting using the internal thoracic artery to prevent coronary ischemia. Moreover, the carotid to subclavian bypass or transposition enables easy access to place a catheter in the left carotid artery for more precise placement of the stent graft in the aortic arch. In doubt, we therefore regard LSA revascularization as indicated.
| Acknowledgments |
|---|
|
|
|---|
| References |
|---|
|
|
|---|
Related Article
This article has been cited by other articles:
![]() |
B. Zipfel, P. Ewert, S. Buz, A. A. El Al, R. Hammerschmidt, and R. Hetzer Endovascular Stent-Graft Repair of Late Pseudoaneurysms After Surgery for Aortic Coarctation Ann. Thorac. Surg., January 1, 2011; 91(1): 85 - 91. [Abstract] [Full Text] [PDF] |
||||
![]() |
I. L. Kron and G. Ailawadi Invited Commentary Ann. Thorac. Surg., August 1, 2009; 88(2): 504 - 505. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |