Long-term Comparative Outcomes of Carotid Artery Stenting Following Previous Carotid Endarterectomy vs De Novo Lesions

Abstract

Introduction

Multiple randomized controlled studies have shown the efficacy of primary carotid endarterectomy (CEA) in reducing the risk of stroke over medical therapy alone.^1,2 However, repeat CEA for recurrent stenosis is technically demanding, with an increased risk for local and neurological complications. Carotid artery stenting (CAS) has evolved into an acceptable alternative to CEA in a select group of patients, and it is not surprising that one of the most common current indications for CAS is a prior ipsilateral CEA.

Restenosis after CEA is likely due to intimal hyperplasia in the first 2 years and progression of atherosclerosis or neoatherosclerosis thereafter.³ Nevertheless, many studies,^4–14 including one from our institution,^13,14 have reported that patch angioplasty closure after CEA decreases the rates of restenosis as well as early occlusion. Although the chance for restenosis is significantly lowered by patch angioplasty, it still occurs. Oszkinis et al¹⁵ reported a 9.3% cumulative incidence of carotid restenosis after CEA during an average 12-month follow-up interval. Also, Goodney et al¹⁶ demonstrated a slight decrease in restenosis, especially in clinically significant restenosis <1 year after CEA, secondary to an increase in patching after conventional CEA and increased attention to patient safety outcomes. However, there are little data available to characterize long-term outcomes of CAS following CEA.

Although some may assume that restenosis is higher following CAS, the CREST (Carotid Revascularization Endarterectomy versus Stent Trial) investigators found rates of restenosis at 2 years were similar between CAS and CEA (6.0% vs 6.3%, respectively; p=0.58) based on a peak systolic velocity (PSV) >300 cm/s duplex criterion correlating to 70% stenosis.^17,18 On the other hand, many respected authorities recommend CAS for restenosis after CEA, and some authors have reported the safety, feasibility, and early postoperative outcomes of CAS after CEA.^19-21 This study offers an extended follow-up comparing de novo CAS with CAS for post-CEA restenosis.

Methods

Results

Patient Groups

The main indication for carotid angiography with possible revascularization was severe carotid stenosis (≥70%-99% on duplex) in both groups (83.6% vs 83.7%, p=0.999). The only difference between the CAS-dn and CAS-res groups was the higher incidence of congestive heart failure in the former (19.4% vs 9.5%, p=0.009). There were proportionally more women in the CAS-res group, but the difference was not significant (45.0% vs 35.2%, p=0.061). For the CAS-res group, the mean elapsed time from CEA to CAS was 72.4±63.6 months. There was no difference in the incidence of PVD between the groups (p=0.248). Symptomatic patients had more renal failure (9.2% vs 3.9%, p=0.044) but did not differ in terms of other clinical variables (Table 1). The time frame between index symptoms and actual stenting in the symptomatic group was 24 to 48 hours.

Table 1.

Patient Characteristics by Study Group and Symptom Status.^a

Variable	CAS-res (n=189)	CAS-d (n=196)	p	Sx (n=207)	Asx (n=178)	p
Age	68.4±9.5	68.9±9.6	0.623	69.0±9.4	68.1±9.8	0.355
Women	85 (45.0)	69 (35.2)	0.061	90 (43.5)	64 (36.0)	0.145
Diabetes	45 (23.8)	56 (28.6)	0.354	53 (25.6)	48 (27.0)	0.728
Renal failure	11 (5.8)	15 (7.7)	0.545	19 (9.2)	7 (3.9)	0.044
Hypertension	103 (54.5)	115 (58.7)	0.534	118 (57.0)	100 (56.2)	1.000
CHF history	18 (9.5)	38 (19.4)	0.009	31 (15.0)	25 (14.0)	0.885
PVD	42 (22.2)	34 (17.3)	0.248	44 (21.3)	32 (18.0)	0.520

Open in a separate window

Abbreviations: Asx, asymptomatic; CAS-dn, carotid artery stenting for de novo disease; CAS-res, carotid artery stenting for restenosis after carotid endarterectomy; CHF, congestive heart failure; PVD, peripheral vascular disease; Sx, symptomatic.

^aContinuous data are presented as the means ± standard deviation; categorical data are given as the counts (percentage).

The majority (94.5%) of patients was treated with a single carotid stent. There were no significant differences in the percentage of patients with postintervention residual stenosis (<30%; 100% each arm) or complications between CAS-res vs CAS-dn: in-hospital stroke (1.4% vs 1.8%, respectively), MI (0.9% vs 0%), or death (0.9% vs 0%).

Mean follow-up was 62.4±45.6 months (median 53.5, range 1–180). Average clinical/TVR follow-up was longer for the CAS-res group (71.9±48.6 months) compared with 53.3±40.5 months for the CAS-dn group (p<0.001). For the ISR outcome, 64 patients had no duplex scans during the 15-year study period. The average ISR follow-up time was longer in the CAS-res group (41.8±33.3 months) compared to the CAS-dn group (27.7±30.9 months, p<0.001). Although non-significant, a higher proportion of ISR was found in the CAS-res group (15.3%) than CAS-dn group (9.6%, p=0.112), as well as more TVRs (7.9% vs. 5.9%, respectively; p=0.452). Ten patients (4 CAS-res and 6 CAS-dn) required a second reintervention.

There were significantly fewer patients presenting with symptoms in the CAS-res (45.8%) vs CAS-dn (59.7%) group (p=0.004). There were more indications for stroke and neck irradiation in the CAS-dn group [9.5 vs 4.7% (p=0.062) and 14.0 vs. 0.9% (p<0.001), respectively; Table 2). However, there was no difference for ISR or TVR in symptomatic compared with asymptomatic patients [ISR: 12.6% vs 12.0% (p=0.875) and TVR: 6.1% vs 7.8% (p=0.571)]. There was no difference in freedom from ISR across the 4 study groups (p=0.174; Figure 1A). Likewise, there was no difference in the freedom from TVR across the study groups (p=0.856; Figure 1B).

Table 2.

Procedure Characteristics by Study Group and Symptom Status.^a

Variable	CAS-res (n=214)	CAS-dn (n=221)	p	Sx (n=230)	Asx (n=205)	p
Indication: stroke	10 (4.7)	21 (9.5)	0.062	31 (13.5)	0 (0.0)	<0.001
Indication: TIA	30 (14.0)	34 (15.4)	0.787	64 (27.8)	0 (0.0)	<0.001
Neck irradiation	2 (0.9)	31 (14.0)	<0.001	20 (8.7)	13 (6.3)	0.371
Contralateral occlusion	4 (1.9)	5 (2.3)	1.000	6 (2.6)	4 (2.0)	0.755
Symptomatic	98 (45.8)	132 (59.7)	0.004	—	—
Severe stenosis ± occlusion by duplex	179 (83.6)	185 (83.7)	1.000	187 (81.3)	177 (86.3)	0.194
Angiographic stenosis	84.6±9.6	83.6±9.7	0.323	83.5±11.1	84.6±7.6	0.262
Vessel diameter, mm	7.4±1.5	7.1±1.2	0.010	7.2±1.3	7.2±1.4	0.758
Lesion length, mm	35.9±7.5	36.9±6.7	0.127	36.2±7.3	36.6±7.0	0.592
Right side intervention	92 (43.0)	100 (45.2)	0.699	100 (43.5)	92 (44.9)	0.773

Open in a separate window

Abbreviations: Asx, asymptomatic; CAS-dn, carotid artery stenting for de novo disease; CAS-res, carotid artery stenting for restenosis after carotid endarterectomy; Sx, symptomatic; TIA: transient ischemic attack.

^aContinuous data are presented as the means ± standard deviation; categorical data are given as the counts (percentage).

Open in a separate window

Figure 1.

Freedom from (A) in-stent restenosis and (B) target vessel reintervention. CAS-dn Asx, carotid artery stenting for asymptomatic patients; CAS-dn Sx, carotid artery stenting for symptomatic patients; CAS-res Asx, carotid artery stenting for asymptomatic patients with previous carotid endarterectomy (CEA); CAS-res Sx, carotid artery stenting for symptomatic patients with previous CEA. The asterisk (*) denotes the time period wherein the standard error exceeds 10%.

During univariate analysis, no medication—either preoperative or discharge—was found to have any significant effect on ISR or TVR (Table 3). However, patients with TVR (vs those without) were more likely to be on a statin either preoperatively (70.0% vs 51.6%, p=0.059) or when discharged (76.7% vs. 59.5%, p=0.081). Multivariate analysis found PVD as the sole independent predictor of ISR (HR 1.92, 95% CI 1.03 to 3.62, p=0.041], while significant predictors for TVR were age at the time of the procedure (<65 years, HR 2.55, 95% CI 1.05 to 6.18, p=0.039) and PVD (HR 2.46, 95% CI 1.03 to 5.87, p=0.043).

Table 3.

Relationships of Medications to Target Vessel Reintervention (TVR) and In-Stent Restenosis (ISR).^a

	TVR			ISR
Medication	No (n=405)	Yes (n=30)	p	No (n=320)	Yes (n=45)	p
Preoperative
Aspirin	310 (76.5)	19 (63.3)	0.122	254 (79.4)	31 (68.9)	0.124
Clopidogrel	246 (60.7)	21 (70.0)	0.340	208 (65.0)	24 (53.3)	0.139
Statin	209 (51.6)	21 (70.0)	0.059	172 (53.8)	24 (53.3)	1.000
Cilostazol	8 (2.0)	1 (3.3)	0.478	7 (2.2)	2 (4.4)	0.307
Discharge
Aspirin	358 (88.4)	25 (83.3)	0.384	285 (89.1)	41 (91.1)	0.802
Clopidogrel	364 (89.9)	27 (90.0)	1.000	289 (90.3)	43 (95.6)	0.403
Statin	241 (59.5)	23 (76.7)	0.081	187 (58.4)	29 (64.4)	0.518
Warfarin	25 (6.2)	0 (0.0)	0.243	15 (4.7)	3 (6.7)	0.475
Cilostazol	11 (2.7)	1 (3.3)	0.581	7 (2.2)	2 (4.4)	0.307

Open in a separate window

^aData are presented as the counts (percentage).

Discussion

Although not the only etiology, extracranial carotid atherosclerosis is a common cause of ischemic stroke and amenable to surgical treatment to prevent an ischemic cerebral event. For symptomatic de novo carotid stenosis, CEA remains the standard of care in our institution. However, CAS may be the preferred treatment for selected patients under certain circumstances. One of the areas in which CAS intuitively makes the most sense is recurrent disease following CEA, but there is limited objective data related to long-term outcomes on that front. In a small series of 27 patients, 8% to 19% restenosis rates were noted after CEA and up to 25% of patients were symptomatic.¹⁹ The Vascular Study Group of New England²⁶ reported that rates for the composite endpoint of stroke/death/MI were statistically similar between repeat CEA vs CAS after prior CEA in both asymptomatic (4.4% vs 3.3%, p=0.8) and symptomatic patients (6.6% vs 5.8%, p=0.9). Also, there was no significant difference in >70% restenosis between repeat CEA and post-CEA CAS groups (5.2% vs 3.0%, p=0.5). However, patients with repeat CEA had increased stroke/death/MI rates in both symptomatic (6.6% vs 2.3%, p=0.05) and asymptomatic (4.4% vs 1.7%, p=0.03) subgroups compared to primary CEA. In addition, a prior ipsilateral CEA was an independent predictor for stroke/death/MI among all patients undergoing CEA [odds ratio (OR) 2.1, 95% CI 1.3 to 3.5].

A similar review from our institution²⁷ with a mean follow-up of 33 months indicated that CAS is as safe as repeat CEA. However, repeat CEA had a higher incidence of transient cranial nerve injury, while CAS had a higher incidence of >50% ISR. Also, Attigah et al²⁰ reported similar midterm overall survival between the two groups. There is a perceived advantage of CAS in recurrent carotid stenosis, including reductions in local and neurological complications in experienced hands, which has promoted significant enthusiasm among multiple authorities for adopting CAS as a valid option for treating recurrent restenosis after ipsilateral CEA.^20,26-30

In terms of symptom status, the ongoing Athero-Express study³¹ indicates that the rate of restenosis is higher in patients treated for asymptomatic atherosclerotic disease compared with more vulnerable plaques. The study quantified the plaque at 2-year periods, including collagen, calcifications, lipid cores, plaque thrombosis, macrophages, smooth muscle cells, and microvessels. Interestingly, large atheroma, plaque thrombosis, macrophages, and calcifications were less frequently observed over time and reflect the current enforcement of risk factor stratification by vascular specialists.

In the current study, patients were followed prospectively with duplex scans at 1, 6, and 12 months and then every 12 months thereafter. Although it did not reach statistical significance, ISR was higher in the CAS-res group, which may warrant close follow-up. Bowser et al³² reported findings similar to our study, as 16% of patients required reinterventions following repeat CEA. We have previously reported a trend toward a higher restenosis rate during follow-up for patients treated with CAS after ipsilateral CEA,²⁴ which persisted in the current study.

It seems intuitive that medications could influence ISR, but this has never held true in carotid disease or other domains. Some researchers have reported that perhaps aspirin use provides a protective effect against ISR.^33–35 The most important lesson from this finding for our group is that we should strive to achieve >95% compliance with aspirin therapy in all our carotid patients. Unexpectedly, under univariate analysis, we found that statin use preoperatively or at discharge was associated with an increased risk for ISR. However, this finding did not hold up under multivariate analysis. The opposite, a protective effect of statin use, has been widely presented by other researchers. Tadros et al³⁶ indicated the protective effect of statin use against embolic particles during CAS (statin: 16.4±2.1 vs no statin: 42.4±9.5, p=0.001). Also, Perler³⁷ in his comprehensive review highlighted pleiotropic effects of statins with 3 times the risk reduction in the rate of perioperative stroke (p<0.05) and a five fold reduction in perioperative mortality (p<0.05) among 1566 patients who underwent CEA. In that report, the advantages of using statins was confirmed in a series of >3000 CEAs performed at multiple Canadian hospitals where statin use was independently associated with a 75% reduction (OR 0.25, 95% CI 0.07 to 0.90) in the odds of death and a 45% reduction (OR 0.55, 95% CI 0.32 to 0.95) in the odds of ischemic stroke or death among patients with symptomatic carotid disease. There could be at least three possible reasons for our finding. First, it could be due to an anomaly in our data and not be a true finding. It is most likely that statin use data could be serving as a marker for more severe disease, and correspondingly, these patients were more likely to develop ISR. On the other hand, our finding could be a true finding, but lacked statistical power under multivariate analysis.

Predictors of restenosis were evaluated in CREST¹⁸; the study identified female gender (HR 1.79, 95% CI 1.25 to 2.56), diabetes (HR 2.31, 95% CI 1.61 to 3.31), and dyslipidemia (HR 2.07, 95% CI 1.01 to 4.26) as independent predictors of restenosis at 2 years. Interestingly, smoking predicted an increased rate of restenosis after CEA (HR 2.26, 95% CI 1.34 to 3.77) but not after CAS (HR 0.77, 95% CI 0.41 to 1.42).¹⁷ In our review, we did not find gender or dyslipidemia to be independent predictors, but only PVD, which was also an independent predictor of TVR along with age <65 years. An important recent meta-analysis³⁸ reported short-term outcomes of 13 studies that included 1132 patients who were treated with CAS or CEA for carotid restenosis after previous CEA. Symptomatic lesions were treated with CAS (30%) or with CEA (40%). After adjusting for all confounders, there was no difference in the endpoint between the groups (2.3% vs 2.7%; adjusted OR 0.8, 95% CI 0.4 to 1.8). Also, over a 13-month median follow-up, there was no difference in the risk of restenosis between the groups, although cranial nerve injuries were more common in CEA patients.

Another large series by New et al³⁹ evaluated 338 patients who underwent CAS with an average interval from CEA of 5.5±7.3 years (39% symptomatic). The authors stated that CAS can be performed safely in patients with restenosis following CEA, with results comparable to most published studies on repeat CEA.

The comparative analysis between CAS-res (n=1996) and CAS-dn (n=10,122) in the CARE (Carotid Artery Revascularization and Endarterectomy) study⁴⁰ utilized propensity score matching based on 28 characteristics to match 1756 patients in each CAS cohort. Authors reported that CAS-res patients had lower procedural adverse event rates and comparable 30-day adverse event rates compared with the CAS-dn group. In addition, other reports have detailed the importance of considering CAS as a low-risk procedure for recurrent carotid stenosis after CEA.^20,24,40

Cuadra et al,⁴¹ in their series of 217 CAS procedures, showed no significant difference between 30-day stroke and stroke + mortality rates between the 2 groups. This study agreed with our report, as there was no significant difference in the composite of death, stroke, or MI at 30 days. To the best of our knowledge, this current study offers the longest follow-up outcomes for patients who had CAS after CEA (Table 4).

Conclusion

The current study provides the first long-term outcome analysis of carotid angioplasty after CEA versus de novo CAS. The study supports previous shorter term or small sample size trials suggesting that CAS is a safe and feasible therapeutic option in patients with recurrent stenosis following CEA. Though nonsignificant, more ISR and TVR were found in patients with a history of CEA. The study results also suggest that PVD is an independent predictor of ISR and TVR, while age at the time of the procedure is an additional predictor of TVR.

Acknowledgments

Footnotes

References