Do device characteristics impact outcome in carotid artery stenting?

Methods 

Patients 

During a 67-month period, from January 2001 until August 2005, 709 patients were scheduled to undergo percutaneous carotid revascularization in the Department of Vascular Surgery of the AZ St-Blasius in Dendermonde, Belgium and the Department of Cardiovascular and Thoracic Surgery of the Imelda Hospital in Bonheiden, Belgium. CAS could not be performed in 8 (1.1%) of 709 patients because of unsuccessful EPD delivery or deployment, and all were converted to carotid endarterectomy. As outlined in Table I, the patient-related, lesion-related, and procedure-related data of the 701 patients receiving CAS (414 men; mean age, 72.4 ± 8.4) were collected and entered into the joined investigational database of the two hospitals for analysis. The assessed variables and their frequency were analyzed for association ≤30 days after CAS with occurrence of stroke, defined as all strokes with symptoms persisting >24 hours; death (all deaths); or transient ischemic attack (TIA), defined as neurologic deficits lasting less than 24 hours.

Table I. Risk factor analysis adverse events within 30 days for the total and symptomatic populations

		Total population (n = 701)			Symptomatic population (n = 301)
	Variable	Frequency, n(%)	OR(95% CI)	P	Frequency, n(%)	OR(95% CI)	P
	Age ≥80	114(16)	2.4(1.0-5.6)	NS	49(16)	2.2(0.6-7.2)	NS
	Left side lesion	328(47)	1.3(0.6-2.9)	NS	125(42)	1.4(0.4-4.2)	NS
	Symptomatic	301(43)	1.6(0.7-3.6)	NS	—	—	—
	Nicotine abuse	179(26)	1.3(0.6-3.1)	NS	79(26)	1.5(0.5-4.9)	NS
	Hypertension	524(75)	0.9(0.4-2.1)	NS	212(70)	1.5(0.3-7.1)	NS
	Diabetes mellitus	158(23)	1.3(0.5-3.1)	NS	68(23)	1.3(0.4-4.5)	NS
	Other PVD	225(32)	1.3(0.5-2.9)	NS	64(21)	2.1(0.7-6.7)	NS
	Hypercholesterolemia	436(62)	2.7(1.0-7.3)	0.0410⁎	178(59)	6.6(0.6-51)	NS
	EPD utilized	671(96)	1.7(0.4-7.4)	NS	276(92)	2.8(0.2-49)	NS
	Lesion predilated	46(6)	1.3(0.3-5.7)	NS	11(4)	0.8(0.04-14)	NS
	Ulcerated Lesion	136(19)	1.3(0.5-3.3)	NS	65(22)	0.9(0.3-3.5)	NS
	Echolucent lesion	480(68)	1.0(0.4-2.4)	NS	191(63)	1.5(0.4-4.8)	NS
	Restenosis after surgery	22(3)	2.4(0.5-11)	NS	5(2)	5.3(0.5-51)	NS
	Open-cell stent used	145(21)	2.2(1.0-5.1)	NS	63(21)	4.1(1.4-12)	0.0136⁎
	Concentric EPD used	119(17)	2.3(0.9-5.4)	NS	48(16)	3.3(1.016-10)	0.0525
	Open-cell stent used + echolucent lesion	110(16)	3.1(1.2-8.2)	0.0343⁎	(N/A)
	Concentric EPD used + echolucent lesion	87(12)	3.7(1.3-10)	0.0174⁎

OR, Odds ratio; CI, confidence interval; PVD, peripheral vascular disease; EPD, embolic protection device; NS, not significant; N/A, not analyzed.

⁎ P = 0.05.

CAS procedure 

CAS was performed according to this unit’s existing standards of care as described previously.20, 21 Protected CAS was performed in 671 patients (95.7%). Distal filtration systems were used in 639 patients (91.2%), proximal occlusion in 31 (4.4%), and distal occlusion in one (0.1%). Eccentric filters (ie, filter eccentrically located of working wire) were selected in 520 patients (74.2%), and concentric filters (ie, filter concentrically positioned on working wire) were administered in 119 (17.0%). Table II gives a detailed overview of the selected filter devices.

Table II. Overview and frequency of used distal filtration embolic protection device

	Type/Name	N (%)
	Eccentric
	FilterWire EX/EZ (Boston Scientific Corp, Natick, Mass)	440(62.8)
	Spider/SpideRX (ev3, Plymouth, Minn)	80(11.4)
	Concentric
	Angioguard XP/RX (Cordis, Miami Lakes, Fla)	74(10.6)
	Emboshield (Abbott Vascular Devices, Redwood City, Calif)	27(3.9)
	Trap (ev3, Plymouth, Minn)	18(2.6)

Carotid stents were used in 695 interventions (99.1%). Closed-cell stents (ie, all stent-struts are interconnected) were used in 549 patients (78.3%) and open-cell stent (ie, not all stent-struts are interconnected) implantation was performed in 145 (20.7%). Table III gives a detailed overview of the selected stents.

Table III. Overview and frequency of carotid stents

	Type/name	N (%)
	Closed cell
	Carotid Wallstent (Boston Scientific Corp, Natick, Mass)	23(74.6)
	X-act (Abbott Vascular Devices, Redwood City, Calif)	23(3.3)
	NexStent (Endotex, Cupertino, Calif)	3(0.4)
	Open cell
	Precise (Cordis, Miami Lakes, Fla)	70(10.0)
	Zilver (William Cook Europe, Bjaeverkov, Denmark)	37(5.3)
	Protégé (ev3, Plymouth, Minn)	31(4.4)
	Memotherm (Bard, Karlsruhe, Germany)	3(0.4)
	Exponent (Medtronic Vascular, Santa Rosa, Calif)	3(0.4)
	Sinus Superflex (Optimed, Ettlingen, Germany)	1(0.1)

Results 

As outlined in Table IV, the overall combined rate of stroke, death, or TIA for the total population within this database was 3.7% (26/701) at 30 days. The breakdown of events at 30 days was TIA, 2.3% (16/701); stroke, 1.0% (7/701); and death, 0.4% (3/701). The rates of events in the symptomatic group were TIA, 3.3% (10/301); stroke, 1.0% (3/301); and death, 0.3% (1/301); and for the asymptomatic group, TIA, 1.5% (6/400); stroke, 1.0% (4/400); and death, 0.5% (2/400).

Table IV. Adverse events at 30 days within the study groups

	30-day outcome	Total % (n = 701)	Symptomatic % (n = 301)	Asymptomatic % (n = 400)
	TIA	2.3(16)	3.3(10)	1.5(6)
	Stroke	1.0(7)	1.0(3)	1.0(4)
	Death	0.4(3)	0.3(1)	0.5(2)
	TIA/stroke/death	3.7(26)	4.6(14)	3.0(12)

TIA, Transient ischemic attack.

ORs for basic risk factors in the total population (age ≥80, left sided lesion, symptomatic, nicotine abuse, hypertension, diabetes mellitus, other peripheral vascular disease, hypercholesterolemia, EPD utilization, predilatation, ulcerated lesion, echolucent plaque, restenosis after surgery) were analyzed and found to be not significant, with the exception of hypercholesterolemia (OR, 2.7; 95% CI, 1.0 to 7.3, P = .041). All hypercholesterolemia patients in the studied population received a dedicated diet program, and 72% received statin treatment. The complete breakdown and analysis of these calculated ORs is summarized in Table I for both the entire population and the symptomatic subgroup.

Subgroup analysis of the 301 symptomatic patients (Table V) showed that in the 298 stented patients, a 30-day combined stroke/death/TIA rate of 11.1% (7/63) was found in patients where an open-cell designed stent was implanted vs 3.0% (7/235) for closed-cell patients, resulting in an OR of 4.1 (95% CI, 1.4 to 12, P = .0136) for open-cell designed stents. In the 280 symptomatic patients protected with a distal filtration system, 30-day combined stroke/death/TIA rates of 10.4% (5/48) and 3.4% (8/232) were recorded for concentric and eccentric filters, respectively. The OR for concentric filters was 3.3 (95% CI, 1.016 to 10, P = .0525) for the matching rate for stroke/death/TIA at 30 days.

Table V. Adverse events at 30 days within the symptomatic CAS patients

	30-day outcome	Stent design		Filter design
		Open cell % (n = 63)	Closed cell (n = 235)	Concentric % (n = 48)	Eccentric % (n = 232)
	TIA	9.5(6)	1.7(4)	8.3(4)	2.6(6)
	Stroke	1.6(1)	0.9(2)	2.1(1)	0.4(1)
	Death	0.0(0)	0.4(1)	0.0(0)	0.4(1)
	TIA/stroke/death	11.1(7)	3.0(7)	10.4(5)	3.4(8)

TIA, Transient ischemic attack.

Subgroup analysis of the 480 patients with echolucent plaques (Table VI) resulted for the 475 stented patients in a 30-day combined rate for stroke/death/TIA of 8.1% (9/110) and 2.2% (8/365) for open-cell vs closed-cell stents. The corresponding OR was 3.1 (95% CI, 1.2 to 8.2, P = .0343) for open-cell stent designs. In case a distal filtration system was selected in the patients with echolucent plaque (n = 430), the 30-day combined stroke/death/TIA rate was 9.2% (8/87) for concentric and 2.0% (7/343) for eccentric filters. This yielded for the 30-day stroke/death/TIA rate an OR for concentric filters of 3.7 (95% CI, 1.3 to 10, P = .0174).

Table VI. Adverse events at 30 days within the CAS patients presenting with echolucent lesions

	30-day outcome	Stent design		Filter design
		Open cell % (n = 110)	Closed cell (n = 365)	Concentric % (n = 87)	Eccentric % (n = 343)
	TIA	7.2(8)	0.8(3)	6.9(6)	1.5(5)
	Stroke	0.9(1)	1.1(4)	1.1(1)	0.3(1)
	Death	0.0(0)	0.3(1)	1.1(1)	0.3(1)
	TIA/stroke/death	8.1(9)	2.2(8)	9.2(8)	2.0(7)

TIA, Transient ischemic attack.

Discussion 

The impact of various risk factors on the 30-day stroke, death, and TIA rate was investigated. No differences could be observed solely considering stroke and death. The statistical significances were found if TIAs were taken into account. Thus, any possible observed advantage or disadvantage was explained by the occurrence of TIAs in the periprocedural period. These temporary neurologic events are likely mediated by small particles that pass through a stent with insufficient scaffolding or through the interstices of a stent deployed in an emboligenic CAS lesion. The observation that most documented events were transient and resolved rapidly argues in favor of these particles being within the small end of the range of those that become clinically symptomatic.

Allowing for this use of TIA in our adverse event rate, our data indicate that in the 298 symptomatic patients who received stents, there appeared to be an elevated OR for adverse events ≤30 days in those treated with an open-cell stent (n = 63) compared with a closed-cell stent (n = 235). Comparison of free cell area indicates that closed-cell stents tend to have a lesser free cell area than open-cell stents.

One analysis evaluated the surface area of various stents used for CAS. The three closed-cell stent designs had cell surface areas of 1.08 mm2 to 4.7 mm2, and the four open-cell designs had cell surface areas of 5.89 mm2 to 11.48 mm2 (Houdart, personal communication). A possible mechanism for closed-cell stent superiority in this series is the low free-cell area in these designs. Indeed the free-cell area of the Carotid Wallstent (Boston Scientific, Natick, Mass) was the lowest of the seven analyzed (1.08 mm2) and was the dominant stent used in this study, having been used in nearly 74% of patients.

Thus, closed-cell stents have an intrinsically greater potential to scaffold and support fractured plaque and support thrombogenic material away from the moving blood pool. These data may indicate that closed-cell stents should be used in symptomatic patients undergoing CAS. Alternative explanations for these observed data effect may exist. For instance, one group has noticed increased platelet activation after open-cell use vs closed-cell stent use.23, 24, 25 More complex hematologic factors may thus contribute to the superiority of closed-cell stent designs in our database.

Likewise, within the 280 symptomatic patients who received filter protection, the OR for protection from adverse events ≤30 days appeared to be elevated with the use of concentric (n = 48) rather than eccentric filters (n = 232). Other clinical and benchtop data suggest that eccentric filters perform superiorly to concentric filters in terms of TIA prevention and particle capture as well as other criteria.26, 27, 28

We postulate the better wall apposition in the distal internal carotid artery (ICA) due to axial flexibility may account for these observations in the use of eccentric filters. The wires of a concentric filter device may pull it away from the wall when used in tortuous anatomy. Filter malposition within the ICA beyond the carotid bifurcation lesion being treated during CAS may, of course, lead to incomplete embolic protection during a given CAS case. It is possible that when in use, floating eccentric filters are more prone to relatively complete wall apposition, thus resulting in somewhat improved embolic protection during CAS when compared with concentric filters used during CAS. We thus believe that eccentric EPDs should be used in symptomatic patients undergoing CAS.

When these ORs were calculated for patients with echolucent plaque, which seems to increase the risk of stroke during CAS,28 similar observations of relative benefit of closed-cell stents and eccentric EPD’s were demonstrated.

Limitations of this study include that it is a retrospective analysis of a centralized database describing a dual-center experience. The database from which these data are extracted is maintained in a registry format and, as such, is subject to the limitations of self-audit. Because these data were collected in two centers closely working together and sharing techniques, it is possible that the protocols may have influence the results observed and thus the conclusions. Retrospective data analysis implies a nonrandomized device selection, which might bias the study outcome if a preselection of devices existed for specific lesion types.

Nevertheless, in both participating centers, carotid procedures at the time were only performed in the perspective of clinical trials and carotid training programs sponsored by different medical device companies. All patients in whom carotid procedures were performed during these single sessions/trials were planned to be treated with, and received, the sponsor’s device. During these sessions, only in exceptional cases where the patients presented with extreme anatomy were the stents were selected accordingly; hence, stents and EPDs were randomly assigned in nearly all of the CAS procedures.

Closed-cell stents and eccentric filter devices predominate among the devices used in this series. It is a potential explanation for our results that a preference by the operator(s) for these devices may be coincident with increased expertise with these devices. This is, however, a high volume center for CAS overall, and experience with even the lesser-used devices in this series is large. Finally, there is a preference within this center for a specific closed-cell stent (eg, Wallstent) and eccentric filter (eg FilterWire, Boston Scientific) for CAS when there is adverse anatomy generally, a bovine left common carotid artery (CCA), a tortuous ICA or CCA, or simply a tougher case overall. If anything, closed-cell stents and eccentric filters saw use in the toughest of the cases in this registry. It stands to reason that closed-cell stents and eccentric filters would be predisposed to perform below other devices if these factors were not controlled for. We did not endeavor to quantify or adjust for technical complexity for each CAS case. Despite this, closed-cell stents and eccentric filters performed better than alternative devices in this analysis.

Optimally, a prospective study comparing closed-cell stent design vs open-cell stents and comparing eccentric vs concentric filters in symptomatic and asymptomatic patients would be useful to address any concerns generated by these observational data. In light of the limitations presented, alternative explanations for these data do exist, and a prospective trial would be necessary before conclusive resolution that the device characteristics under consideration here materially impact patient outcome.

Conclusions 

Our data support the preferential use of closed-cell stents and eccentric EPDs in patients undergoing CAS in whom echolucency is documented on preprocedure duplex examination or who are symptomatic. Further analysis of device design variables may be warranted. Particularly in symptomatic patients or those with echolucent lesions, closed-cell design and eccentric filters seem superior in the short term. Prospective investigation comparing open-cell vs closed-cell stents and eccentric vs concentric filter devices might be valuable in further addressing these questions. Until further data become available, strong consideration should be given to use of closed-cell stents and eccentric EPDs in symptomatic patients or patients with echolucent lesions when other factors do not mandate use of other platforms.27

Author contributions