Author + information
- Received March 25, 2010
- Revision received September 16, 2010
- Accepted September 16, 2010
- Published online February 8, 2011.
- Paola De Rango, MD⁎,⁎ (, )
- Gianbattista Parlani, MD⁎,
- Fabio Verzini, MD, PhD⁎,
- Giuseppe Giordano, MD⁎,
- Giuseppe Panuccio, MD⁎,
- Matteo Barbante, MD⁎ and
- Piergiorgio Cao, MD†
- ↵⁎Reprint requests and correspondence:
Dr. Paola De Rango, Unit of Vascular and Endovascular Surgery, University of Perugia, Hospital S. M. Misericordia, Piazza Menghini 1, 06134 Perugia, Italy
Objectives This study sought to evaluate long-term outcomes of carotid stenting (CAS) versus carotid endarterectomy (CEA) based on physician-guided indications.
Background The issue regarding long-term outcome of CAS versus CEA in patients with carotid stenosis is clinically relevant but remains unsettled.
Methods Consecutive patients (71% men, mean age 71.3 years) treated by CEA (n = 1,118) or CAS (n = 1,084) after a training phase were reviewed. Selection of treatment was based on better-suitability characteristics (morphology and clinical). Data were adjusted with propensity score analysis and stratified by symptoms, age, and sex.
Results Thirty-day stroke/death rates were similar: 2.8% in CAS and 2.0% in CEA (p = 0.27). The risk was higher in symptomatic (3.5%) versus asymptomatic (2.0%) patients (p = 0.04) but without significant difference between CAS and CEA groups. Five-year survival rates were 82.0% in CAS and 87.7% in CEA (p = 0.05). Kaplan-Meier estimates of the composite of any periprocedural stroke/death and ipsilateral stroke at 5 years after the procedure were similar in all patients (4.7% vs. 3.7%; p = 0.4) and the subgroups of symptomatic (8.7% vs. 4.9%; p = 0.7) and asymptomatic (2.5% vs. 3.3%; p = 0.2) patients in CEA versus CAS, respectively. Cox analysis, adjusted by propensity score, identified statin treatment (p = 0.016) and symptomatic disease (p = 0.003) associated with the composite end point. There were no sex- or age-related significant outcome differences.
Conclusions When physicians use their clinical judgment to select the appropriate technique for carotid revascularization CAS can offer efficacy and durability comparable to CEA with benefits persisting at 5 years.
Durable stroke prevention is the objective of treatment of carotid stenosis. Although several randomized controlled trials (RCTs) have addressed the issue of outcomes of carotid stenting (CAS) versus carotid endarterectomy (CEA) (1–4), many of these trials suffered from lack of interventional operator experience or inconsistent use of embolic protection devices and the quality of CAS procedures was low (2–4). Furthermore, in all RCTs, the results were available at <4 years after CAS/CEA, and the information on CAS is now affected by the lack of substantial data on delayed outcomes. The question on what is the long-term outcome of CAS compared with CEA in patients with a carotid stenosis is clinically relevant and remains unsettled. The purpose of this study was to compare the 5-year safety of CEA with CAS performed with cerebral protection and current techniques in a large population outside RCTs by using physician-guided assignment of treatment.
Patients entered into a prospectively compiled computerized database of all primary extracranial carotid revascularizations (CAS and CEA) performed at a single vascular surgical center from January 2001 to March 2009 were analyzed. All patients with either >60% symptomatic or >70% asymptomatic carotid stenosis were treated by surgeons. For the purpose of the study, patients who received revascularization for recurrent carotid stenosis and bypass grafts were excluded. All procedures were performed after the training phase. The revascularization treatment choice (CAS/CEA) was left to the discretion of the treating surgeon according to better suitability and periprocedural risk evaluation as well as plaque and vessel morphology and presence of comorbidities. This assignment strategy was used in other nonrandomized studies (5). Usually, patients with unfavorable aortic arch anatomy, severe peripheral vascular disease precluding femoral access, or extremely tortuous carotid anatomy were excluded from CAS. Similarly, known allergies to aspirin, clopidogrel, or contrast media and renal insufficiency were considered exclusion criteria for CAS. High-neck carotid bifurcation and long carotid lesions as well as obesity and ongoing dual antiplatelet therapy were relative contraindications for CEA.
In our center, the caseload necessary to perform safe CAS (assuming 2% as the safety threshold per year for major periprocedural complication rate) included the first 195 procedures. It was after this training that the rate of disabling strokes (mainly occurring during the catheterization and filter time frames of CAS) significantly decreased and remained stable at <2% per year in each of the following years (6). To avoid bias due to the learning curve effect of the operators, the first 195 CAS performed within the training phase (2001 to 2003 interval) were excluded from this study. With increasing experience and by applying the best suitability morphology and clinical criteria learned during the training, the number of CAS by year expanded to reach and overcome the numbers of CEA. In the more recent years, CEA was selected for fewer and usually the most complex cases (e.g., acute symptoms, unstable plaque). Therefore, CEAs performed in the last 2 years of the study (2008 and 2009, when these high-risk selection criteria were used) were excluded from the present analysis to avoid possible major selection bias. Of carotid revascularization procedures performed by year for primary carotid stenosis, 50% in 2004, 50.5% in 2005, 19.5% in 2006, and 19.7% in 2007 were performed by CEA.
All patients received aspirin (100 to 325 mg once daily) and clopidogrel (75 mg once daily) or ticlopidine (250 mg twice daily) before the procedure. Carotid stenting was carried out following a standardized protocol in an endovascular room equipped with a high-quality fixed-imaging system (Axiom Artis FA, Siemens, Berlin, Germany).
Minimal or no sedation was used during the procedure and neurological status was continuously monitored. Intravenous heparin (100 U/kg) was routinely given before selective catheterization of the common carotid artery and not reversed at the end of the procedure. Percutaneous transfemoral or transbrachial approaches under local anesthesia were used to allow selective engagement of the target carotid artery. Variable models of cerebral protection devices and carotid stents (open cell, close cell, or hybrid configuration; tapered or straight) were employed in all procedures. The choice of specific material depended on vessel anatomy and lesion characteristics. Close cell stents were generally used in straight vessels and in the presence of soft plaque. There was a higher preference for using a filter as cerebral protection; in selected cases (carotid tortuosity, extremely soft plaque profile at ultrasound, and so on), crossing a lesion with a filter was avoided by using a proximal occlusion system. The reference vessel diameter was determined (stent size/length choice) according to pre-operative ultrasound measurements. In the case of a large discrepancy between common carotid and internal carotid artery diameters, a tapered stent was deployed. After stent deployment, completion angiography was performed. Closure devices for the access control were used since 2006.
Post-procedure antiplatelet therapy included dual drug treatment (aspirin and thienopyridines) for a minimum of 1 month and continued at the treating physicians' discretion.
Carotid endarterectomy was performed under local or general anesthesia. All patients were under antiplatelet 1-drug therapy, either aspirin (100 to 325 mg once daily) or ticlopidine (250 mg twice daily) or thienopyridine (clopidogrel 75 mg daily). Shunts were used selectively according to clamping intolerance. Eversion or patch and occasionally direct primary closure were used as arterial closure techniques. Patient monitoring followed the same modality as CAS in awake patients; stump pressure measurements were used when general anesthesia was employed. Systemic heparinization was used at the same dosage as CAS and then reversed after carotid declamping.
With coronary artery disease, we included patients with documented history of angina or myocardial infarction, regardless of duration and type of treatment received. Patients with peripheral artery disease were considered those with documented lower limb intermittent claudication or critical limb ischemia. Features and time of pre-operative symptoms were evaluated by external neurological audit. Patients were defined as symptomatic when ipsilateral hemispheric or retinal symptoms occurred within 6 months from the procedure. Stroke was defined as any new hemispheric or retinal neurological event persisting >24 h and classified as fatal, disabling (modified Rankin score ≥3), or nondisabling (modified Rankin score <3). Myocardial infarction was diagnosed by the cardiologist in the occurrence of persistent ST-segment changes and/or new Q-wave in 2 leads or the presence of elevated enzymes (including troponin >0.1 ng/ml).
The degree and characteristics of carotid stenosis at baseline and during follow-up were assessed with duplex ultrasound by experienced operators who defined site, degree, length of stenosis, plaque characteristics, and vessel measurements previously validated against angiography as a gold standard technique. “Complex carotid plaque” was judged by ultrasound when lack of uniform pattern and prevalence of soft appearance was evident. Data were confirmed by grossly intraoperative assessment during CEA. Degree of stenosis was defined according to the University of Washington modified duplex velocity criteria (7). Contrast-enhanced computed tomography was performed selectively, in cases of uncertainty at ultrasound examination. Degree of stenosis was always confirmed by visual evaluation with angiography during CAS.
Cerebral computed tomography scan was used in symptomatic patients to assess the extent of recent lesions if any.
Patients scheduled for CAS or CEA with antiplatelet intolerance or under anticoagulation for coexisting medical comorbidities continued to receive their baseline therapy. Written consent was obtained from all patients before revascularization.
Primary end point was the combined risk of any stroke or death within 30 days (perioperative) and any ipsilateral stroke after the procedure. Secondary end points were stroke, death, transient ischemic attack (TIA), myocardial infarction (MI), and local complications (hematoma, cranial nerve injuries) occurring within 30 days and the rate of stroke, death, and restenosis after the procedure.
Outpatient clinical and ultrasound examinations were scheduled at regular intervals (at 6 and 12 months and yearly thereafter) and symptom status was assessed. Carotid restenosis was set at >50% using ultrasound criteria (7). Repeat computed tomography angiography or angiography was performed only if repeat revascularization was being considered. Patients were instructed to report any new neurological symptoms occurring after hospital discharge. In cases of neurological symptoms or uncertainty occurring anytime after the procedure, the patients were evaluated by a certified independent neurologist expert in vascular disease.
Analysis of data was by treatment actually received. Data are shown as frequencies and percentages for categorical variables and mean (±SD) for continuous variables. Characteristics of patients in CAS and CEA groups were compared using chi-square and Fisher exact tests (when appropriate) for categorical variables, and analysis of variance or Student t test for continuous variables. Unadjusted and adjusted odds ratios (ORs) with corresponding 95% confidence intervals (CIs) were used to compare outcomes between CAS and CEA.
Analyses of perioperative and late outcomes were stratified by pre-operative symptoms to account for evident differences in symptomatic and asymptomatic distribution of patients.
The rates of end points at 5 years were estimated with Kaplan-Meier method to compensate for patient dropouts and the level of significance was calculated with log-rank test and its standard error (SE). Curves were displayed up to a value of SE <0.10. Symptomatic and asymptomatic patients were analyzed separately.
Because surgery or endovascular treatment was not randomly assigned in this population, potential confounding and selection biases were addressed by analyzing the rate of the composite outcome (any stroke or death within 30 days and any ipsilateral stroke after procedure) with multivariate analyses after using backward elimination methods. The following variables were included in the model: treatment (carotid stenting), age, sex, pre-operative symptoms, contralateral occlusion, coronary disease, peripheral artery disease, diabetes, hypertension, statin therapy, and complex plaque. Interactions among the 11 covariates and composite outcome were assessed with Cox regression analysis.
To address the differences between CAS and non-CAS patients in a nonrandomized study, a propensity score was also derived, reflecting the probability that a patient would undergo CAS. This was accomplished by multivariable logistic regression models using CAS as the dependent outcome variable and entering all measured baseline patients characteristics shown in Table 1 (except for bilateral procedure and atrial fibrillation) as covariates (8,9). The score was used in trying to adjust the estimates of the main composite outcome using 3 approaches with Cox regression models: unadjusted, adjusted for raw propensity score, and adjusted for propensity score and other covariates selected in multivariate models using stepwise selection method (9).
Subgroup analyses in models stratified by symptoms, sex, and age were performed to balance for potential subgroup discrepancies in CAS versus CEA. For age categories, we used the age cutoff applied in the CREST (Carotid Revascularization Endarterectomy Versus Stenting Trial) (≥69 years vs. ≥70 years; >80 years).
A value of p < 0.05 was considered statistically significant for all measurements. Stat-Calc-EPIINFO 6.0 (PC version 3.5.1, Centers for Disease Control and Prevention [CDC] Atlanta, Georgia) and SPSS (PC version 13.00 Win package, SPSS, Inc., Chicago, Illinois) were used for all data analyses.
Over the study period, 2,202 interventions for primary carotid stenosis were performed in 2,041 patients: 1,084 CAS (in 1,007 patients) and 1,118 CEA (in 1,034 patients). There were 1,562 men and 640 women; mean age was 71.3 years (range 46 to 92 years).
Demographic and baseline characteristics for CAS and CEA populations are displayed in Table 1.
The CAS patients were less likely to have a history of peripheral artery disease or symptomatic disease and more likely to have hypertension, coronary disease, diabetes, and be on statin treatment for hyperlipidemia (Table 1). Complex carotid plaque, as judged by ultrasound, was more frequently detected in CEA (41.9% vs. 31.2%, p = 0.001).
Carotid stenting was originally planned in 2,004 procedures. In 20, the procedure was not accomplished because it was not possible to approach or cross the carotid lesion for the following reasons: difficulty in cannulation of bovine arch (n = 2), extreme vessel tortuosity (n = 10), pre-occlusive lesions (n = 6), or both pre-occlusive lesion and tortuosity (n = 2). These patients were censored from CAS and analyzed according to the treatment they actually received. The remaining 1,084 completed CAS procedures were all accounted for the CAS group. In this group, 4 patients required stent removal for acute complications after successful CAS. Three male patients (2 with both atrial fibrillation and anticoagulant therapy) developed acute stent thromboses associated with TIA immediately after an uneventful procedure. In a female patient, repeated TIA occurred after stent deployment for plaque protrusion through the stent struts. All the 4 patients were uneventfully converted to CEA. There was no technical failure in the CEA group.
The 30 day (periprocedural) risk of stroke or death in overall CAS and CEA populations was 2.5% (54 of 2,202); 18 periprocedural strokes were disabling.
Periprocedural outcome measures in CAS compared with CEA are reported in Table 2. There were no significant differences in periprocedural risk of stroke or death between CAS and CEA procedures: 2.8% (31 of 1,084) after CAS versus 2.0% (23 of 1,118) after CEA; p = 0.27 (Table 2). There were no deaths in the CAS group. In the CEA group, 3 fatal strokes and 3 cardiac deaths occurred. One patient developed severe congestive heart failure and died 3 days after CEA performed for a recent stroke. Another patient developed fatal MI after she was uneventfully discharged. The third cardiac death occurred in a patient who developed neck hematoma, MI, and pneumonia soon after surgery and died on post-operative day 4.
Any perioperative major adverse events (including any stroke, death, and any MI) occurred in 3.1% of CAS and 2.7% of CEA (p = 0.61). The rates of periprocedural MI were similarly distributed between CAS and CEA patients: 0.3% in CAS versus 0.4% in CEA (p = 1.00). Rates of TIA were higher in CAS (3.6%) than in CEA (1.1%) patients, p < 0.001, whereas cranial nerve injuries were higher in CEA (4.4%) than in CAS patients (0%; p < 0.001).
Among the combined CAS and CEA population (n = 2,202), the periprocedural stroke or death rate was lower in asymptomatic patients, 2.0% (30 of 1,518), than in symptomatic patients, 3.5% (24 of 684; p = 0.04). Among symptomatic patients (n = 684), the rate was 4.5% in CAS versus 2.9% in CEA, p = 0.29; among asymptomatic patients (n = 1,518), the rate was 2.3% in CAS versus 1.6% in CEA; p=0.36 (Table 3).
Distribution of periprocedural stroke or death risk between CAS and CEA in women and in men and within the ≤69-, ≥70-, and >80-years age categories of patients are shown in Table 3.
Mean follow-up was 33.05 ± 21.7 months. All the patients, with the exception of 59 unavailable, had a minimum of 6 months of follow-up. Follow-up ranged from 6 to 108.4 months.
During the observation period, 178 patients died (n = 74 in CAS, n = 104 in CEA) and 12 cerebral hemorrhages (1 nonfatal) were recorded (n = 5 after CAS, n = 7 after CEA). The 5-year survival rates from any cause mortality were similar for CAS versus CEA (82.0% vs. 87.7%, p = 0.050) populations (Fig. 1).
The actuarial incidence of ipsilateral late stroke after the procedure was 0.9% in CAS versus 2.7% in CEA. There were no differences at 5 years after the procedure in Kaplan-Meier estimates of the major composite end point (combined risk of any stroke or death within 30 days and any ipsilateral stroke after the procedure) between CAS (3.7%) and CEA (4.7%) groups (p = 0.4) (Fig. 2).
There were no significant differences between CAS and CEA in Kaplan-Meier composite end point rates for any subgroup (symptomatic, asymptomatic, women, men, older, or younger patients) comparisons. Among symptomatic patients, the rates were 4.9% in CAS versus 8.7% in CEA (p = 0.67); among asymptomatic patients, the rates were 3.3% in CAS versus 2.5% in CEA (p = 0.2). Among women, the rates were 4.2% in CAS versus 6.4% in CEA (p = 0.3); among men, the rates were 3.6% in CAS versus 4.0% in CEA (p = 0.07). In the subgroup analysis by age, among older patients (>70 years) rates were 4.0% in CAS and 5.4% in CEA (p = 0.4); among younger patients (≤69 years) rates were 3.2% in CAS and 3.5% in CEA (p = 0.9).
During follow-up, recurrent stenosis of 50% or more was detected in 66 patients (28 CAS and 38 CEA) without significant difference between CAS and CEA patients in Kaplan-Meier estimates at 5 years: 3.4% versus 5.8% (p = 0.7) (Fig. 3). Only 6 recurrent stenoses (4 in CEA and 2 in CAS) led to neurological symptoms.
Cox regression analysis after adjusting for 11 potential confounders with backward elimination demonstrated that symptomatic stenosis (hazard ratio [HR]: 2.15; 95% CI: 1.32 to 3.5; p = 0.002) was the only significant positive predictor of the composite end point (combined risk of any stroke or death within 30 days and any ipsilateral stroke after the procedure). The use of pre-operative statin was the only negative predictor (HR: 0.46; 95% CI: 0.25 to 0.83; p = 0.01) associated with composite end point.
Propensity for CAS use
Patients had a higher propensity to be receiving CAS with a history of coronary artery disease (OR: 1.5; p < 0.001), hypertension (OR: 1.2; p = 0.041), diabetes (OR: 1.2; p = 0.018), hyperlipidemia (OR: 1.2, p = 0.039), statin therapy (OR: 1.33, p = 0.007), and increasing age (OR: 1.33, p = 0.007). Patients had a lower propensity to be receiving CAS with a history of neurological symptoms (OR: 0.59, p < 0.001), peripheral disease (OR: 0.49, p < 0.001), and presence of complex plaque (OR: 0.61, p < 0.001). After adjusting for the derived propensity score, the associations with the main outcome remained essentially unchanged. Whether the propensity score was allowed to be removed or was forced into the Cox model (either with or without other covariates) did not influence the final prediction significantly. Carotid stenting was not an independent predictor of the main composite clinical end point in this analysis (HR: 1.4; 95% CI: 0.85 to 2.33; p = 0.17). Statin was confirmed as a negative independent predictor (HR: 0.49; 95% CI: 0.27 to 0.87; p = 0.016) and symptomatic carotid as a positive predictor of the composite end point (HR: 2.0; 95% CI: 1.2 to 3.26; p = 0.003).
Both CAS and CEA appear to have low perioperative complications and excellent longer-term results and may be considered useful tools for preventing stroke in patients with carotid stenosis. In trained settings using modern technology and appropriate physician-guided indications of best suitability, catheter-based treatment with stent for carotid stenosis can be successfully accomplished with low periprocedural major complication (stroke/death rate: 2.8%) and can provide durable resolution of carotid stenosis in 96.6% of patients at 5 years of follow-up with low risk of late ipsilateral stroke.
Our periprocedural and late data, based on physician-guided and not on random assignment of treatment, are similar to those recently shown by the CREST study, which has been the most recent and largest RCT (over 2,500 patients enrolled), comparing CAS (n = 1,262) and CEA (n = 1,240) in average risk patients, and is the only randomized trial to include asymptomatic and symptomatic patients and to require only well-trained operators for both the procedures before randomization (1). Our results confirmed and estimated at 5 years the initial findings of the CREST study (assessed at a median follow-up of about 2.5 years after treatment): the main composite end point (combined risk of any stroke or death within 30 days and any ipsilateral stroke after the procedure) rates were 3.7% in CAS versus 4.7% in CEA (p = 0.4). Nevertheless, our data offer the opportunity to analyze outcomes in a large number of consecutive patients treated for carotid stenosis with CAS and CEA reflecting the modern “real-world” scenario outside the selected within-trial population.
In addition, this study provides data on direct comparison between CAS and CEA in the longer-term showing that the nonprocedural risk of stroke might be particularly low after CAS: 0.9% at 5 years. This finding, in accordance with other literature data (10–14), is important because some observers had anticipated that an increased rate of late stroke after CAS could be expected because of a most likely higher probability of restenosis.
Although the aim of any treatment of carotid stenosis is long-term prevention of stroke, the long-term rate of neurological events after carotid angioplasty with stenting (by using today's technique) remains uncertain because little data, and none from RCTs, on late clinical outcome after CAS (14–17) are available. Although, without exception, the EVA-3S (Endarterectomy Versus Angioplasty in Patients With Symptomatic Severe Carotid Stenosis) (11), the SAPPHIRE (Stenting and Angioplasty with Protection in Patients at High Risk for Endarterectomy) (12), the SPACE (Secondary Prevention With Antioxidants of Cardiovascular Disease in End-Stage Renal Disease: Randomized Placebo-Controlled Trial) (10), and the CREST (1) studies, each reported that following successful stenting (i.e., assuming no strokes or deaths in the first 30 days) CAS was as durable as CEA with high freedom from late stroke rates, the length of follow-up in these studies was restricted to a maximum of <4 years.
The SPACE trial (10) suggested that carotid restenosis risk might be higher after endovascular treatment: life-table estimates of restenosis up to 2 years were 10.7% in the stenting versus 4.6% in the CEA groups (p = 0.0009), even though only 2 restenoses were symptomatic. Conversely, our data showed that the risk of restenosis after CAS might be slightly lower than after CEA: 3.4% versus 5.8% (p = 0.7) at 5 years, respectively. Most restenoses after CAS occurred early, likely as a consequence of intimal hyperplasia or healing process, and very few were associated with recurrence of symptoms. Nevertheless, criteria for stenosis measurements, different thresholds, and ultrasound-subjective assessments make comparisons in restenosis rate among different studies unreliable (18,19). “Restenosis” should be regarded as a secondary, minor end point in studies evaluating stroke prevention.
It has been also suggested that in CAS patients the apparent small long-term risks might be decompensated by the higher periprocedural risk, defeating any benefit provided from revascularization. Patient selection is essential to assign best-suited patients to the lower-risk treatment. Periprocedural risk of nondisabling strokes might be higher during CAS and cardiac risk might be higher during CEA (1), as our data in part confirmed. The lack of randomization and the use of physician-guided more than random assignment of treatment allowed in our population a larger prevalence of patients with cardiac disease in the CAS arm (36.7% vs. 26.6%; p < 0.001). Yet, CAS patients compared with CEA patients showed no increase (0.3% vs. 0.4%) in periprocedural cardiac events (MI), whereas all 3 cardiac deaths occurred in the CEA group, confirming CEA as a procedure with higher cardiac hazards. Our symptomatic patients showed higher risks than asymptomatic patients in either the periprocedural period or the long term. For symptomatic patients, the risk of periprocedural stroke during CAS was higher although there were no statistically significant differences in rates when compared with CEA patients (4.5% vs. 2.9%; p = 0.29).
Our study failed to show any significant increase in early and late risks for CAS versus CEA in women or in older patients. Nevertheless, like the CREST trial (1), a trend toward increased periprocedural risk in ≥70-years patients (3.6% vs. 2.0%, p = 1.05) and decreased risk in younger patients (1.6% vs. 2.0%, p = 0.8) was found.
Based on this study, patients with unfavorable aortic arch anatomy, extreme tortuous carotid anatomy, or severe peripheral vascular disease precluding femoral access, as well as known allergies to aspirin or clopidogrel should be considered preferentially for CEA. Patients with severe coronary disease, high-neck carotid bifurcation, and obesity should be considered at increased risk during CEA and might be best suited for CAS. The use of statin was the only variable to be independently associated with improved long-term outcome.
Our data reflect practice in a dedicated high-volume vascular center including large-volume endovascular/vascular surgeons who benefit from the cooperation and support of a team of interventionalists and neurologists. Our findings might not be reproducible under less ideal conditions (and may not reflect management of patients treated at many community hospitals). Second, patients were not randomized; this study rather reflects a retrospective observation and results may partially be related to selection bias. Even after adjustment with the aid of multivariate models, propensity score, and subgroup stratifications, the full clinical picture that determines outcomes in surgery versus endovascular therapy may not be fully balanced. Finally, as with all retrospective studies, the database is subject to referral and ascertainment bias, including patient reliability in accurately reporting new symptoms that could allow investigators to underestimate neurological event rates assigned by independent neurologists in both CAS and CEA groups.
Our results demonstrate that using clinical and morphology criteria we were able to select CAS rather than CEA for patients undergoing carotid revascularization without compromising long-term vessel patency and stroke-free survival. The periprocedural risks, similar to those of CEA, and durability of symptom-free outcomes at 5 years support the efficacy of CAS as a reasonable alternative to CEA in experienced centers.
The authors thank Ms. Francesca Zannetti and Ms. Eileen Mahoney for language and editing assistance.
The authors have reported that they have no relationships to disclose.
- Abbreviations and Acronyms
- carotid stenting
- carotid endarterectomy
- myocardial infarction
- randomized controlled trial
- transient ischemic attack
- Received March 25, 2010.
- Revision received September 16, 2010.
- Accepted September 16, 2010.
- American College of Cardiology Foundation
- Feit F.,
- Brooks M.M.,
- Sopko G.,
- et al.
- Faught W.E.,
- Mattos M.A.,
- van Bemmelen P.S.,
- et al.
- D'Agostino R.B. Jr.
- Wu A.H.,
- Aaronson K.D.,
- Bolling S.F.,
- Pagani F.D.,
- Welch K.,
- Koelling T.M.
- Cremonesi A.,
- Gieowarsingh S.,
- Spagnolo B.,
- et al.
- Randall M.S.,
- McKevitt F.M.,
- Kumar S.,
- et al.
- Nederkoorn P.J.,
- Brown M.M.