Endovascular management of spontaneous carotid artery dissection

Methods 

Procedure 

All patients were pretreated with clopidogrel (75 mg by mouth daily for at least 4 days before the procedure or a 600-mg oral load if emergently treated) and aspirin (325 mg by mouth daily). Patients underwent four-vessel digital subtraction angiography. Intravenous heparin was administered to achieve an activated clotting time of longer than 250 seconds. All patients were treated under local anesthesia. Stable vessel access was achieved by placing a sheath or guide catheter distally in the common carotid artery. The true lumen of the dissected vessel was identified by using high-resolution angiography with high frame rates (5-10 frames per second) and multiple projections. It was often helpful to perform nonsubtracted angiograms to better visualize the thin luminal flaps, which can be subtracted out with digital subtraction techniques. When there was doubt about the true lumen, a 0.014-inch floppy-tipped coronary microwire was passed distal to the lesion into the presumed true lumen; a microcatheter or 1.5-mm balloon catheter was then inserted distal to the lesion, and the wire was removed. If there was no blood return, the catheter or balloon was withdrawn. If there was blood return, a very gentle manual injection of 0.25 to 1 mL of contrast was injected through the microcatheter or central balloon lumen to confirm that the tip was in the true lumen. When this was confirmed, the wire was replaced, and the procedure was continued. If the catheter tip was in a false channel, then the catheter was withdrawn, and wiring of the true lumen was reattempted. We used the softest wire possible and paid particular attention to good wiring technique to avoid extending the dissection flap and entering pseudoaneurysms, given the risk of perforation.

All lesions causing luminal narrowing were predilated with a small (2.5- to 4-mm) coronary balloon. A combination of self-expanding or balloon-expandable stents was used in each vessel treated, depending on the vessel size and location: intracranially extending lesions (eg, petrous carotid segment) were treated with balloon-expandable stents, whereas, because of their flexible nature, cervical carotid lesions were treated with self-expanding stents when possible. Stent grafts were used in all cases involving expanding pseudoaneurysms. The external carotid artery orifice was not covered during these procedures (Fig 1).

View Large Image Download to PowerPoint

Figure. A, Left intracranial and extracranial internal artery before intervention. B, Left intracranial internal carotid artery after intervention. C, Left extracranial internal carotid artery after intervention.

Results 

Patient characteristics 

Seven patients (5 women) between the ages of 39 and 57 years (mean ± SD 47.7 ± 2.8 years) were treated (Table I). Five of the seven patients had a stroke before treatment as a result of the initial arterial dissection. The time interval between the initial dissection and the first endovascular procedure is listed in Table I. Six patients had angiographic evidence of dissection in two vessels, and one patient, in three vessels. The indications for endovascular treatment included ongoing ischemia of 12 to 24 hours’ duration (despite anticoagulation with heparin) in three patients, expanding pseudoaneurysms in three, and severe reduction in cerebral blood flow in one. One patient had 3 vessels treated, 3 patients had 2 vessels treated, and 3 patients had 1 vessel treated, for a total of 12 treated vessels. Of the patients with two or more treated vessels, two had simultaneous treatment of both internal carotid arteries (ICA). One of these patients was comatose from bihemispheric ischemia, and the other had recurrent strokes and bihemispheric emboli on transcranial Doppler monitoring despite heparin and aspirin therapy. A third patient had the left ICA treated because of recurrent stroke and had the right internal and common carotid arteries treated subsequently because of severe flow-limiting stenosis involving both vessels. The location of the dissections was the right ICA (n = 6), the right common carotid artery (n = 1), and the left ICA (n = 5).

Table I. Clinical characteristics

	Patient No.	Age (y)	Sex	Presenting symptoms	Time fromfirst SCD tofirst stent	FMD?	Infarct location	Vessels with dissection	Indication for treatment
	1	41	F	Bifrontal headache, neck pain, tinnitus in R ear, and numbness of R side of face and arm	2y	Y	R hemisphere watershed infarcts	RICA LICA	Expanding pseudoaneurysm (expanding over 3 mo) with recurrent TIA and intractable headaches
	2	57	M	Syncope, R homonymous hemianopia, aphasia, R hemiparesis, and obtundation	8d	N	L parietal-occipital	LICA RICA	Highly compromised bihemispheric cerebral blood flow
	3	39	F	L Horner syndrome, tinnitus, and headache	1y	Y	None	RICA LICA L Vertebral	Severe flow-limiting dissection and expanding pseudoaneurysm (expanding over 1 y)
	4	53	M	R-sided neck pain and severe recalcitrant headache	5mo	N	None	RICA L Vertebral	Expanding pseudoaneurysm (expanding over 4 mo) with intractable headaches
	5	40	F	L lower extremity numbness	5d	Y	R insular cortex and internal capsule	RICA LICA	Ongoing bihemispheric ischemia (and emboli by TCD) despite heparin and aspirin
	6	54	F	Right upper extremity weakness and mild expressive aphasia	3mo	Y	L frontal-parietal	RICA RCCA LICA	Compromised cerebral blood flow
	7	50	F	Right frontal headache, L hemianopsia, dysarthria, and L hemianesthesia	5d	Y	R frontal-temporal	RICA LICA	Compromised cerebral blood flow with fluctuating clinical deficits

SCD, Spontaneous carotid artery dissection; L, left; R, right; ICA, internal carotid artery; TIA, transient ischemic attack; CCA, common carotid artery; FMD, fibromuscular dysplasia; TCD, transcranial Doppler ultrasonography.

Table II lists the procedural details. Self-expanding stents were used in 11 vessels. The single balloon-expandable stent was placed in a relatively immobile segment: the distal cervical ICA extending into the petrous segment. A total of four stent grafts were placed in the three segments that had pseudoaneurysms; one patient (patient 1) required overlapping Jomed (Jomed International AB, Helsingborg, Sweden) stents to completely cover a long pseudoaneurysm. Angiographic success was achieved in all patients, with no significant residual stenoses or filling of the pseudoaneurysms.

Table II. Summary of angiographic characteristics and procedural details by vessel treated

	Vessel treated	Balloons and stents	Pretreatment/posttreatment degree of stenosis	Complications
	L ICA and pseudoaneurysm	4 × 30-mm Maverick⁎ balloon 5 × 26-mm Jomed† stent #1 5 × 26-mm Jomed stent #2 5 × 20-mm Viatrac‡ balloon	50% → 0%, pseudoaneurysm completely occluded	None
	R ICA	4 × 40-mm Viatrac balloon 7 × 40-mm Precise§ stent 5 × 20-mm Viatrac balloon	99% → 0%	Hemorrhagic conversion 13 d after intervention
	L ICA	4 × 40-mm Viatrac balloon 6 × 40-mm Precise stent	70% → 0%	None
	R ICA	4 × 13-mm Tetra‡ stent	80% → 5%	None
	L ICA pseudoaneurysm	5 × 19-mm Jomed stent 5 × 20-mm Viatrac balloon	50% → 0%, pseudoaneurysm completely occluded	None
	R ICA and 3 pseudoaneurysms	8 × 30-mm Wallgraft⁎ 7 × 40-mm OptaPro§ balloon	80% → 0%, pseudoaneurysms completely occluded	Small, non–flow-limiting, intimal tear R CCA
	R ICA	7 × 20-mm Precise stent	90% → 10%	None
	L ICA	7 × 40-mm Precise stent 5 × 30-mm Viatrac balloon	90% → 0%	None
	L ICA	6 × 30-mm Precise stent 5 × 20-mm Viatrac balloon	70% → 0%	None
	R CCA extending into ICA	7 × 8-mm SMART§ stent 5.5 × 30-mm Aviator§ balloon 7 × 20-mm Precise stent 5.5 × 30-mm Aviator balloon	90% → 0%	None
	R ICA	1.5 × 15-mm Maverick balloon 5 × 20-mm Precise stent 5 × 40-mm Precise stent 4.5 × 20-mm OpenSail‡ balloon	99% → 0%	None

L, Left; R, right; ICA, internal carotid artery; CCA, common carotid artery.

⁎ Boston Scientific Inc (Boston, Mass). † Jomed Inc (Helsingborg, Sweden). ‡ Guidant Corp (Indianapolis, IN). § Cordis Corp (Miami Lakes, FL).

Outcomes 

There were no immediate clinical complications, and the three patients with ongoing clinical deficits had immediate improvement or resolution of recurrent transient ischemic attacks (Table III). There was one small, non–flow-limiting, intraprocedural dissection that was asymptomatic. The patient with bihemispheric ischemia and obtundation developed hemorrhagic conversion of a large infarct 13 days after the intervention but has since had continued recovery and is living independently. One patient died after heart transplantation (4 months after stenting).

Table III. Immediate and long-term outcomes

	Patient No.	Immediate outcome	Long-term outcome	Most recent anatomic follow-up
	1	No new symptoms	Resolution of tinnitus, improvement in headache, no recurrent ischemia	L ICA: 12-mo A/G patent
	2	Awakened, R hemiparesis resolved	Hemorrhagic conversion 13 d after intervention, with recovery; stable neurologic examination at 2 y after intervention	R and L ICA: 24-mo U/S patent
	3	No new symptoms	Resolution of Horner syndrome and tinnitus, improvement in headache severity and frequency	R and L ICA: 24-mo A/G patent
	4	No new symptoms	Resolution of intractable headaches	R ICA: 14-mo U/S patent
	5	Transient ischemic attacks and TCD emboli resolved; placed on transplant list	Stable neurologic examination; died as a complication of transplantation 4 mo after intervention	R and L ICA: 3-mo U/S patent
	6	No new symptoms, subtle R upper extremity weakness	Symptoms resolved	L ICA and R ICA/CCA: 12-mo U/S patent
	7	Immediate improvement in hemianesthesia and hemianopsia	Minor residual visual field and sensory deficits	R ICA: 6-mo U/S patent

R, Right; L, left; ICA, internal carotid artery; A/G, angiography; U/S, ultrasonography; CCA, common carotid artery; TCD, transcranial Doppler ultrasonography.

All of the vessels were patent on follow-up imaging (catheter angiography, carotid duplex scan, or CTA; Table III). The average duration of follow-up was nearly 14 ± 8 (SD) months (range, 3 to 24 months). At their most recent follow-up visits, all patients showed improvement or resolution of their previous neurologic deficits, and none had recurrent ischemia. The patients with headache had either complete resolution of headaches or marked improvement to the point that they no longer required headache-specific therapy.

Discussion 

This case series demonstrates that SCD can be safely and effectively treated with stenting: it is the largest published series to date of patients treated with this technique. All of the patients in this series were treated with stents to maintain vessel patency by using a technique similar to that used for atherosclerotic carotid stenosis.16 However, we did not use emboli-prevention devices because we believed that the risk of worsening the dissection or creating a new intimal tear was greater than the potential for embolization given the lack of bulky plaque underlying the stenoses. In all of our patients, there was improvement in the symptoms. In addition, three of our patients had ongoing cerebral ischemia that resolved immediately after improvement of cerebral blood flow, with rapid return of neurologic function. This improvement was attributed to a return of penumbral tissue to normal or near-normal function, the removal of the embolic source, or both. None of these patients received thrombolysis.

There have been 8 previously reported series of patients with carotid dissections (both traumatic and spontaneous) treated with endovascular stent angioplasty (a total of 57 patients with 22 spontaneous dissections). In these series, no deaths occurred, no transient ischemic attacks or strokes occurred that could be attributed to the treated vessel, and patients remained clinically stable or improved during the follow-up period.6, 7, 8, 9, 12, 13, 14, 15

Only a minority of patients with SCD require such treatment. More than 85% of medically treated SCD patients improved clinically and angiographically in one study.5 The standard of care for SCD is anticoagulation with warfarin or antiplatelet therapy. Even patients in whom the dissection is severe and persistent have a low risk of recurrent stroke if the dissection is unilateral.17 Studies using magnetic resonance imaging to follow up on changes in arteries with spontaneous dissections also find stabilization or improvement in most cases.18, 19 In some circumstances, however, medical therapy is inadequate (eg, patients with recurrent symptoms despite anticoagulation,6, 7, 8 patients with expanding or symptomatic pseudoaneurysm,7, 9 and patients with involvement of multiple vessels or poor collateral vessels8) or contraindicated (eg, patients in whom anticoagulation is contraindicated because of a risk of bleeding6, 7). Multivessel involvement increases the risk of compromised cerebral blood flow8 and, in this series, may explain why our patients all had evidence of multivessel involvement. In addition, there is a group of patients in whom the degree of stenosis does not improve or actually worsens. This may be accompanied by additional symptoms, especially when more than one vessel is involved.20

In the above-mentioned groups and in patients who experience additional symptoms while fully anticoagulated, there is growing evidence that an intervention should be considered. Surgical intervention has been used for the repair of spontaneous dissections, but it is technically challenging and is associated with higher morbidity and mortality than the repair of atherosclerotic lesions of the carotid.21 In one recent series, the surgical treatment of SCD in 49 patients was associated with a 12% incidence of stroke and death and a 58% incidence of cranial neuropathy.3 No examples of stroke or death associated with endovascular treatment of SCD have been reported in the literature. In our series, there were no immediate neurologic complications, although one of our patients had hemorrhagic conversion of a large infarct 13 days after the intervention. This may have been related to the tendency of large bland infarcts to develop hemorrhage, particularly in the setting of dual treatment with aspirin and clopidogrel. A second possibility is that the increased cerebral blood flow after endovascular repair and impaired autoregulation caused the rupture of an ischemic vessel wall, ie, hyperperfusion syndrome. This same patient had been nearly comatose before stenting as a result of severe bihemispheric ischemia and is currently alive and living at home. Endovascular treatment may therefore be the treatment of choice for patients with persistently symptomatic SCD,4 but because the natural history of most cases of SCD is so good, it is unlikely that most patients with SCD would benefit from endovascular therapy.

The clinical significance of a persistent pseudoaneurysm is uncertain. El-Sabrout and Cooley10 looked at 12 previously reported series and found that pseudoaneurysms had a combined major stroke and death rate of 21% when treated nonoperatively, whereas Touzé et al22 followed up 33 patients with pseudoaneurysms for 14 months and found no evidence of embolic events or rupture. Nevertheless, all of our patients with pseudoaneurysms were symptomatic: most had recalcitrant headaches, and one had recurrent cerebral ischemia. Several reports have used expanding or symptomatic pseudoaneurysms as a criterion for intervention.7, 9 We used stent grafts to exclude and prevent expansion of the pseudoaneurysms and to reconstruct a normal lumen. In addition to technical success, this approach led to resolution of symptoms. Others have described stenting with noncovered stents followed by coil embolization for the treatment of pseudoaneurysms.23 Unfortunately, this approach has been associated with aneurysm recurrence and is more time-consuming.14 Covered stent placement is relatively straightforward, although the selection of covered stents is somewhat limited. The use of the Jomed stent requires manufacturer approval because the stent is available for use only in coronary arteries on a humanitarian device exemption issued by the Food and Drug Administration.

We have clinical and imaging follow-up data for all of our patients. All of the patients have remained symptom free. The method of radiologic follow-up was selected on a case-by-case basis. Stents placed in the proximal cervical ICA or distal common carotid artery were well visualized by ultrasonography. More distal cervical and intracranial lesions cannot be visualized by ultrasonography, although both carotid duplex scan and transcranial Doppler ultrasonography can be useful because they can show a proximal high-resistance pattern (duplex scan) or a distal low-resistance pattern (transcranial Doppler ultrasonography) in cases of significant in-stent stenosis. The utility of magnetic resonance angiography as a follow-up technique for the detection of in-stent stenosis is limited by metal artifact caused by the stent and by the difficulty in visualizing the in-stent segment.24 Similarly, CTA is limited by metal and bone artifact.25 In such patients, catheter angiography is the most sensitive means of detecting in-stent stenosis. We did perform angiography in the patients with distal stents and did not find any evidence of restenosis or dissection in any of the patients. Restenosis may not be a major issue in this group of patients, because their stenoses were nonatherosclerotic. Even the patients with stent grafts have not had any in-stent stenosis or recurrence of pseudoaneurysms. Endovascular treatment therefore seems durable, and it is not clear that follow-up catheter angiography is absolutely necessary if patients are asymptomatic, but, again, this must be decided on a case-by-case basis.

The major limitations of this study are the small number of patients and the fact that it was a retrospective study. These limitations notwithstanding, our experience highlights the potential use of endovascular therapy in selected patients.

Conclusions 

Our study is the largest series to date of patients with SCDs treated with endovascular stenting. It has provided additional evidence for the safety and efficacy of this treatment when medical therapy is insufficient or inappropriate.