Outcomes of Carotid Angioplasty and Stenting for Radiation-Associated Stenosis

Methods

Since November 1996, 16 patients with a history of head and/or neck malignancy treated with radiation therapy were referred to our department for diagnosis and treatment of carotid occlusive disease (Table 1). Institutional review board approval was obtained for the study and written informed consent was obtained for the procedure.

All patients underwent a diagnostic cerebral angiogram on a dedicated biplanar neuroangiographic unit (Neurostar; Siemens, Munich, Germany) to document high-grade carotid stenosis (>70% according to NASCET) in either the internal carotid artery (ICA) or common carotid artery (CCA). Transfemoral access was used for all patients. Patients were premedicated with aspirin 325 mg orally each day 5 days before the procedure through 1999 (9 procedures). All subsequent procedures were performed by using clopidogrel bisulfate 75 mg (Plavix; Bristol-Myers Squibb/Sanofi Pharmaceuticals Partnership, NY) orally each day in addition to aspirin. The activated clotting time was maintained between 250 and 300 seconds with heparin during the procedure.

Following a diagnostic cerebral angiogram, the diagnostic catheter was exchanged for a 6-, 7-, or 8F Shuttle sheath (Cook, Bloomington, IN) placed in the carotid artery proximal to the stenosis. In patient 15, the Shuttle sheath could not be advanced because of the acuity of the angle of the proximal left CCA from the aortic arch and because the proximal location of the common carotid stenosis prevented purchase for the exchange. Therefore, a 5F HN4 catheter (Cook) was used to select the left CCA. Through this catheter, a Prowler Plus microcatheter (Cordis Corp., Miami Lakes, FL) and a Transcend 14 guidewire (Boston Scientific, Natick, MA) were used to catheterize and cross the left carotid stenosis. The Transcend 14 was then replaced by a 0.018-inch exchange length guidewire (Roadrunner; Cook) over which the angioplasty and stent placement were performed after removal of the HN4 and Prowler catheters. Intravenous atropine or glycopyrrolate (American Regent, Shirley, NY) was administered to the 14 patients who had not undergone prior ipsilateral carotid endarterectomy for prophylaxis of bradycardia.

In all but 3 patients, balloon angioplasty was performed before stent deployment. A total of 22 self-expanding stents were used: 9 Smart stents (Cordis), 6 Precise stents (Cordis), and 7 Wallstents (Boston Scientific). In patient 7, tortuosity in the proximal left CCA prevented placement of a Smart stent, so a balloon-expandable coronary stent (NIR; Boston Scientific) was used. Except for patient 5’s right ICA (stented with a Smart stent), patients 1–6 (Table 2, in chronological order) received a Wallstent for their first or only procedure (last Wallstent used in May 1999). The Smart or Precise stent was adapted in lieu of the Wallstent because their cell characteristics allowed for abrupt changes in stent lumen diameter. The stents were oversized by 1–2 mm to the largest diameter of the artery to be treated. Thirteen stent placements were followed by balloon angioplasty. Poststent angioplasty was not performed in vessels that were sufficiently dilated with initial angioplasty and stent placement. No cerebral protection device was used. Intracranial images were routinely obtained before and after the intervention, to confirm patency of intracranial vessels.

Femoral hemostasis was obtained by using 4 Perclose devices (Abbott Laboratories, Abbott Park, IL) in 4 procedures on 3 patients and by using 2 AngioSeal devices (St. Jude Medical, St. Paul, MN) in 2 procedures on one patient. Manual compression was used in the remaining 17 procedures after the activated clotting time had returned to normal. Patients were monitored in the intensive care unit following the procedure with mean arterial pressure maintained within 20 mm Hg above or below baseline. They were discharged the following day, with the exception of the one patient, who suffered a periprocedural stroke. Aspirin and Plavix were continued for 30 days following the procedure, and aspirin was continued for life.

The patients’ medical records were obtained from their referring physicians. Follow-up consisted of periodic carotid Doppler studies, repeat cerebral angiograms, CT angiography (CTA), or clinical history and physical examinations.

Results

Of the 16 patients, 5 were women and 11 were men, with ages ranging from 52 to 79 years (mean, 65 years). The interval from radiation therapy to carotid stent placement ranged from 1 to 23 years. Eleven of 16 (69%) patients had neurologic symptoms consistent with carotid disease: 9 transient ischemic attacks (TIAs) and 2 strokes (Table 1). Three patients (5, 10, and 11) had bilateral carotid stenoses that were treated separately. The first vessels were treated for symptomatic lesions followed (after 2, 4, and 1 month, respectively) by treatment of asymptomatic lesions on the contralateral side, making 19 treated vessels (14 ICAs and 5 CCAs). In the total 19 stented vessels, 23 procedures were performed: 22 stent placement procedures and 1 repeat angioplasty (Table 2). The mean stenosis was 84% (range, 70%–99%). One patient was treated with 2 stents (patient 15). One patient had 30% residual stenosis (patient 3), with the remaining procedures leaving minimal or no residual stenosis.

The procedural stroke rate was 1/23 (4%). Patient 4 developed a small left parietal infarct immediately after the procedure because of rupture of a poststent angioplasty balloon despite inflation pressure below the maximum recommendations (Fig 1). Immediate repeat left carotid angiography showed a widely patent left CCA and ICA with no intracranial emboli. It was not known whether foreign material or other embolus was responsible. The patient was transferred to a rehabilitation facility 13 days postprocedure with mild weakness of the right arm and leg along with mild speech difficulty, which resolved completely over the following month. The procedural TIA rate was 0/23 (0%). Patient 3 developed a 14-cm postprocedure right groin hematoma that resolved without treatment or transfusion. There were no other procedural complications, including no retroperitoneal hemorrhages, infections, or cranial nerve palsies. The 30-day postprocedure complication rate was 0/23 (0%); no neurologic symptoms were reported.

• Download figure
• Open in new tab
• Download powerpoint

Fig 1.

A case of long-term stent patency in patient 4, a 68-year-old woman with thyroidectomy for carcinoma followed by postoperative radiation, bilateral radical neck dissection, and permanent tracheostomy, who experienced transient ischemic changes in the left cerebral hemisphere 21 years after radiation therapy. The patient had a history of graft placement between her left CCA and left subclavian arteries (5 years before admission) as well as a right CCA vein graft bypass (10 years before admission).

A, Digital subtraction angiography (DSA) reveals a 70% stenosis (arrow) of the distal left CCA. Note the bypass graft (arrowhead) between the left CCA and left subclavian arteries.

B, Successful stent placement with little residual stenosis. The proximal end of the 8 × 40 mm Wallstent (Boston Scientific, Natick, MA) extends across the origin of the left subclavian graft (arrow). Poststent angioplasty with an 8 × 40 mm Olbert balloon (Boston Scientific) resulted in rupture of the balloon.

C, Repeat DSA 4.5 years after stent placement shows a patent carotid artery with interval auto expansion of the stent.

The mean follow-up was 28 months (median, 16 months; range, 5–78 months). Of the 19 vessels stented, 15 (79%) remained patent (<50% stenosis) throughout the follow-up period (Table 3). Two of 19 (11%) vessels had repeat angioplasty and stent placement. Most the recurrent stenoses have not been at the initial site of angioplasty but rather at the proximal or distal ends of the stent. Patient 5 developed a new 70% stenosis at the distal margin of the first stent by 13 months. This has remained patent for 50 months following the repeat procedure. Patient 2 required 2 repeat angioplasty and stent placement procedures (Fig 2). The first repeat procedure was done at 61 months for a new 70% stenosis at the proximal margin of the original stent. A third procedure 17 months later resulted in a third stent placed for a new 75% stenosis at the proximal margin of the second stent. The third stent also traversed a 75% in-stent stenosis. One of 19 (5%) had a repeat angioplasty with minimal residual stenosis without the placement of a second stent (patient 12). As in patient 5, this was also done for a new stenosis at the distal margin of the first stent. Repeat procedures were performed when >70% restenosis was found on routine imaging, except for patient 2’s first repeat procedure, which was performed for recurrent TIA at 61 months. In summary, of the 4 treated restenoses, there were 2 at the distal stent margin, one at the proximal stent margin, and one at the proximal stent margin with concurrent in-stent stenosis.

• Download figure
• Open in new tab
• Download powerpoint

Fig 2.

A case of multiple restenting in patient 2, a 61-year-old man with bilateral radical neck dissection and external beam radiation for laryngeal cancer 15 years before right cerebral hemisphere stroke due to right ICA occlusion.

A, DSA shows an 85% ulcerated stenosis (arrow) of the proximal left ICA.

B, DSA after angioplasty and stent placement shows little residual stenosis. Prestent angioplasty was performed with a 4 × 20 mm Symmetry balloon (Boston Scientific). This was followed by placement of an 8 × 40 mm Wallstent (Boston Scientific). Poststent angioplasty was performed by using an 8 × 40 mm Olbert balloon (Boston Scientific).

C, Repeat DSA 5 years after stent placement shows 70% restenosis (arrow) in the distal left CCA at the inferior margin of the stent. The external carotid artery no longer fills.

D, DSA after angioplasty and stent placement with a 5 × 20 mm Savvy balloon (Cordis Corp., Miami Lakes, FL) followed by a 7 × 20 mm Precise stent (Cordis). There is minimal filling of the external carotid artery.

E, DSA 17 months later shows a focal high-grade in-stent restenosis (arrow) and a less severe stenosis in the CCA (arrowhead).

F, Immediate DSA following placement of a third stent (7 × 40 mm Precise; Cordis) and poststent angioplasty with a 6 × 40 Savvy balloon (Cordis) shows a good angiographic result.

Two of 19 (11%) vessels occluded per Doppler examination 14 and 22 months postprocedure (patients 2 and 7). Before documented occlusion, patient 7 had a left hemisphere stroke at 14 months shortly after developing atrial fibrillation. Two patients (one carotid artery in each patient [1 and 14]) were lost to follow-up. Patient 1 died of metastatic adenocarcinoma 12 months after the stent placement of her right ICA. Aside from patients 2 and 7, noted above, no new neurologic symptoms were reported during the follow-up period. All patients were followed with periodic carotid Doppler studies, repeat cerebral angiograms, or CTA, with the exception of patients 9 and 15. These 2 were followed clinically, because imaging studies were not performed to confirm patency.

Discussion

The effects of radiation on normal tissue were investigated at the time of its initial widespread therapeutic use in the early 1940s. Warren discussed extensively the cardiovascular changes associated with radiation (11). Extensive animal studies have demonstrated that radiation induces changes consistent with atherosclerosis: intimal proliferation, selective disruption of the internal elastic lamina, and necrosis in both small and large vessels (12–14).

Many studies in humans have demonstrated increased prevalence of carotid occlusive disease in patients who have received local radiation therapy (15–20). Most recently, a study of 40 patients who had received cervical radiation for head and neck cancer found that 16 (40%) patients had significant (>50%) carotid artery stenosis (20). Six of these patients had unilateral complete carotid occlusion, and 6 had significant bilateral carotid stenosis. Three had sustained a stroke after radiation therapy. Another study, by Elerding et al, of 910 patients surviving at least 5 years after neck irradiation for Hodgkin disease, non-Hodgkin lymphoma, or primary head and neck neoplasm found the incidence of stroke was 6.9% (63 patients) during a mean observation period of 9 years (21). The expected number of strokes per 1000 people in the general population over the same time period is 38 (22). These studies demonstrate that the prevalence of carotid disease in this population is clinically significant and requires follow-up to determine if and when intervention becomes necessary.

The benefits of carotid endarterectomy in general were demonstrated by NASCET. In the case of asymptomatic disease, the Asymptomatic Carotid Atherosclerosis Study (ACAS) estimated the 5-year risk of ipsilateral stroke and any perioperative stroke or death in good surgical candidates with carotid artery stenosis of 60% or greater was 5.1% versus 11.0% in medically treated patients. This study is the basis for surgical treatment of asymptomatic disease in good surgical candidates (23).

In the case of radiation-associated disease, a series of 24 patients treated surgically has been reported (24). In this study, 20 of the 24 patients underwent standard carotid endarterectomy, 2 underwent external carotid endarterectomy, one underwent aortocarotid bypass, and one had a saphenous vein bypass. Following the procedure, 6 patients demonstrated transient cranial nerve palsies, one patient had a postoperative hematoma, which did not require evacuation, and two patients suffered postoperative wound infections. Follow-up ranged from 1 to 156 months, with only 6 of the 24 patients having a follow-up period longer than 18 months. No patients suffered any neurologic events during their follow-up. Duplex scan examination documented one occlusion 1 month postprocedure and one restenosis that required a second operation. The remainder of the grafts were widely patent. Although this study did not show a high rate of procedural complications, it is acknowledged that in this population of patients, radiation-induced scarring and previous neck dissections complicate the surgical treatment.

Of the 16 patients in the present study, 11 had prior radical neck dissections, and 4 had prior CEA. A recent study comparing 124 repeat CEAs with 265 primary CEAs has shown that repeat CEA carries a higher rate of perioperative stroke (4.8% vs 0.8%; P = .015) and cranial nerve injury (17% vs 5.3%; P < .001) (25). The 4 patients in this study with prior CEAs had an uncomplicated periprocedure course with patent vessels through their follow-up period.

Our results have been echoed in three recent studies. In a series of 14 patients, one patient suffered a periprocedural stroke but recovered completely in 2 days (26). There were no other procedural complications. At 6 months postprocedure, nine (64%) patients demonstrated <50% stenosis on carotid Doppler studies, and no patients had any neurologic events. At 18 months, 3 patients had died of unrelated causes, and the others remained asymptomatic. Another study reported 7 patients with 10 treated vessels (27). Six of 7 patients were symptomatic; 4 of 7 had undergone previous radical neck dissection. There was no evidence of early restenosis after a mean follow-up of 8 months. Most recently, in a series of 16 patients who underwent 17 procedures, one postprocedure stroke occurred as a result of thromboembolism to the ipsilateral middle cerebral artery and led to hospital death; there was also one TIA (28). There were 3 recurrent stenoses on duplex scan and one repeat angioplasty with no new neurologic events over a median 30-month (range, 5–55 months) follow-up.

The procedures in this study were performed before the widespread availability of cerebral protection devices. The rate of ischemic complications, however, was low (4% procedural, 0% at 30 days) in these postradiation, symptomatic and asymptomatic, NASCET-excluded patients. Fifteen of 19 vessels (79%) were widely patent as demonstrated by Doppler examination, angiography, or CTA, and the patients remained free of new neurologic symptoms at the end of a mean follow-up period of 28 months.

The benefit of using cerebral protection devices has not been demonstrated in this population, and their use is not without complications. It may not be possible to cross a high-grade stenosis with a filter or balloon-type protection device, and doing so may result in complete occlusion or dissection. In these instances, a nonlesion-crossing balloon device such as the Parodi antiembolism system (W. L. Gore and Associates, Newark, DE) may be of value. Future studies are necessary to show the impact of such devices on the procedural complication rate in this subgroup of patients.

A limitation of the study is its small size (23 procedures and 16 patients). Also, in any study of head and neck cancer patients, another difficulty is obtaining long-term follow-up in view of their comorbidities. In our series, one patient died of her malignancy before follow-up was obtained.

Conclusion

In light of the evolution of self-expandable stents, the use of carotid angioplasty and stent placement has increased rapidly. Although it is clear that endarterectomy is preferred to medical management, the benefit of endovascular stent placement versus endarterectomy remains unclear in the general population. In the population of patients treated with neck irradiation who were excluded from NASCET, angioplasty and stent placement may be preferred in light of the difficulty of surgical management. In this population, the procedure was safe and effective with few complications and good midterm results.