Abstract
Carotid artery stenting (CAS) has emerged as a potential alternative to carotid endarterectomy (CEA) for the management of carotid artery stenosis. The purpose of this article is to provide an evaluation and critical overview of the trials comparing the early and later results of CAS with CEA for symptomatic carotid stenosis. The Cochrane Controlled Trials Register, PubMed/Medline, and EMBASE databases were searched up to February 1, 2009, to identify trials comparing the long-term outcomes of CAS with CEA. The MeSH terms used were “carotid artery stenting,” “carotid endarterectomy,” “symptomatic carotid artery stenosis,” “treatment,” “clinical trial,” “randomized,” and “long-term results,” in various combinations. One single-center and three multicenter randomized studies reporting their long-term results from the comparison of CAS with CEA for symptomatic carotid stenosis were identified. All four studies independently reached the conclusion that CAS may not provide results equivalent to those of CEA for the management of symptomatic carotid stenosis. A higher incidence of recurrent stenosis and peri- and postprocedural events accounted for the inferior results reported for CAS compared with CEA. Current data from randomized studies indicate that CAS provides inferior long-term results compared with CEA for the management of symptomatic carotid artery stenosis. However, it can be argued that all of these trials were performed when both CAS equipment and CAS operators had not evolved to their current status. Given that current equipment and mature experience are required for CAS before comparing it with the current “gold standard” procedure (CEA), the results of soon-to-be reported trials (Carotid Revascularization Endarterectomy vs Stenting Trial [CREST], International Carotid Stenting Study [ICSS], or others) may alter the current impression that CAS is inferior to CEA for the treatment of symptomatic carotid stenosis.
Keywords
In the last few years, carotid artery stenting (CAS) has emerged as a possible alternative to carotid endarterectomy (CEA) for the management of carotid artery stenosis. 1,2 The supporters of CAS underline the less invasive nature of the procedure, which enables its application in high-risk patients, the shorter in-hospital stay, and the reduction in wound complications and cranial nerve injury. 1,2 In contrast, the supporters of CEA dispute the inappropriateness of CEA for high-risk patients 3 and emphasize the disadvantages of CAS, namely, the increased equipment and hospital cost and its not-yet-established long-term effectiveness. 4
According to the recent clinical practice guidelines of the Society for Vascular Surgery, CEA plus optimal medical therapy is the “gold standard” for symptomatic patients with moderate to severe (> 50%) carotid stenosis, as well as for asymptomatic patients with moderate to severe (≥ 60%) carotid stenosis and a low perioperative risk; both are strong recommendations based on high-quality evidence. 5 CAS is suggested as a potential alternative to CEA in symptomatic patients with moderate to severe (≥ 50%) carotid stenosis and a high perioperative surgical risk; this, however, is a weak recommendation based on low-quality evidence. 5 The same guidelines recommend against the use of CAS for asymptomatic patients with moderate to severe (≥ 50%) carotid artery stenosis; this is a strong recommendation but is based, however, on low-quality evidence. 5
In the past few months, several new findings have emerged that may raise questions about the long-term effectiveness of CAS for the management of symptomatic carotid artery stenosis. Asymptomatic carotid stenosis seems to be a process that behaves differently (eg, “stable” versus “unstable” [emboli-producing] carotid plaque for asymptomatic versus symptomatic carotid stenosis) 6 and that will not be dealt with further in this article. This article presents an evaluation and critical overview of the evidence regarding the place of CAS in the management of symptomatic carotid artery stenosis.
Methods
The Cochrane Controlled Trials Register, PubMed/Medline, and EMBASE databases were searched up to February 1, 2009, to identify randomized trials comparing the early and later outcomes of CAS with CEA for patients with symptomatic carotid artery stenosis. The MeSH terms and key words used were “carotid artery stenting,” “carotid endarterectomy,” “symptomatic carotid artery stenosis,” “treatment,” “clinical trial,” “randomized,” and “long-term results,” in various combinations. The reference lists of the gathered reports were manually searched. Additional reports were also considered.
Results
One single-center 7 and three multicenter randomized studies 8–13 comparing the late results of CAS with CEA for the management of symptomatic carotid artery stenosis were identified (Table 1). A brief description of the findings of these studies is presented.
Randomized Trials Comparing the Long-Term Results of Carotid Endarterectomy with Carotid Artery Stenting for the Management of Symptomatic Carotid Artery Stenosis
CAS = carotid artery stenting; CAVATAS = Carotid and Vertebral Artery Transluminal Angioplasty Study; CEA = carotid endarterectomy; CI = confidence interval; EVA-3S = Endarterectomy versus Angioplasty in Patients with Symptomatic Severe Carotid Stenosis; HR = hazard ratio; SPACE = Stent-Protected Percutaneous Angioplasty of the Carotid vs. Endarterectomy.
*For endovascular patients treated successfully (n = 213), stents were used in 55 (26%) and balloon angioplasty alone in 158 (74%).
Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVATAS)
The first multicenter randomized trial comparing the late results of surgical with endovascular treatment for symptomatic carotid stenosis was CAVATAS. 8,9 In CAVATAS, 504 patients with carotid artery stenosis were randomized to CEA or endovascular treatment (253 vs 251 patients, respectively). Stents were used in only about a quarter of the patients receiving endovascular treatment (26%); the remaining three-quarters (74%) received balloon angioplasty alone. Most patients included in CAVATAS were symptomatic (amaurosis fugax, transient ischemic attack [TIA], or stroke; 242 vs 246 patients, or 96% vs 97%, for CEA vs CAS, respectively). The mean carotid artery stenosis was similar in the two groups (86.4% ± 9.1% vs 85.1% ± 10.2% for CAS vs CEA groups, respectively). After a mean follow-up of 1.95 (interquartile range 1.0–2.2) versus 1.98 (interquartile range 1.0–2.8) years for CAS versus CEA, respectively, the rates of death or disabling stroke in any vascular territory and the rate of any ipsilateral stroke did not differ between the groups (14.3% vs 14.2% for CAS vs CEA, respectively; hazard ratio [HR] 1.03; 95% confidence interval [CI] 0.64–1.64; p = .9). At 1 year, severe (70–99%) recurrent carotid stenosis on ultrasonography was more common in patients undergoing endovascular treatment compared with CEA (25 vs 7 patients, or 14% vs 4%, respectively, p < .001). 8,9 Furthermore, in the subgroup of patients treated by stenting, severe stenosis or occlusion at 1 year was found in 9 of 41 patients (22%) who had ultrasound follow-up.
Steinbauer and Colleagues
The higher ultrasound restenosis rates for CAS compared with CEA were recently confirmed in one single-center randomized study. 7 In this small prospective study, the long-term results of 43 patients undergoing CAS were compared with those of 44 patients undergoing CEA (mean follow-up: 66 ± 14.2 vs 64 ± 12.1 months for CAS vs CEA, respectively). 7 All patients had > 70% symptomatic carotid artery stenosis (ie, TIA, amaurosis fugax, or stroke). At 1 year and compared with CEA, CAS patients experienced more episodes of ipsilateral stroke (0 of 44 vs 1 of 43, respectively), TIA (2 of 44 vs 3 of 43, respectively), and recurrent stenosis > 70% (0 of 44 vs 2 of 43, respectively). 7
Final data were available for 32 CAS patients and 29 CEA patients. 7 Two CEA patients and 1 CAS patient were lost to follow-up, whereas 10 CAS and 13 CEA patients died. It was demonstrated that fewer patients undergoing CEA experienced stroke episodes (0 of 42 vs 4 of 42, respectively; p < .05), had > 70% restenosis (0 of 29 vs 6 of 32, respectively; p < .05), and had to undergo reintervention for > 70% restenosis (0 of 29 vs 5 of 32, respectively; p < .05) compared with CAS. 7 Of the ultrasonographically detected recurrent carotid stenoses found in the CAS arm, only two were diagnosed at 7 and 12 months after stent implantation; the four others were detected at 1.7, 2.2, 5.9, and 6.3 years after CAS, respectively. 7 Reintervention and surgical stent removal were necessary in five of the six CAS cases with recurrent stenosis. The remaining patient presented with an ipsilateral stroke 1.3 years after CAS, had a medium-grade stenosis (50–70%), and refused further intervention. 7
This study had several drawbacks: (1) it was a small, single-center study; thus, the results should be interpreted with caution; (2) CAS was performed without the use of embolic protection devices; and (3) the recruitment period was relatively early (August 1999–April 2002). 14 Therefore, the stents used were older and possibly less well suited to CAS than the ones used currently, and the interventionalists possibly did not have enough experience with CAS. 14 In support of this, a recent report presenting an analysis of the periprocedural complications of CAS underlined the importance of appropriate and considerable experience before undertaking systematic use of CAS. 15
Endarterectomy versus Angioplasty in Patients with Symptomatic Severe Carotid Stenosis (EVA-3S)
EVA-3S began in November 2000 to evaluate whether CAS is not inferior to CEA with regard to procedural risks and long-term efficacy in patients with ≥ 60% symptomatic carotid artery stenosis. 10 All patients had suffered a hemispheric or retinal TIA, retinal infarct (amaurosis fugax), or nondisabling stroke within 120 days before enrolment. 10 It was calculated that 872 patients would have to be enrolled to have a statistical power of 80% to assess whether CAS is not inferior to CEA with regard to the 30-day incidence of stroke or death. In September 2005, the safety committee recommended that enrolment in the trial be stopped for reasons of safety and futility. 10 Up to that point, 527 patients had been randomly assigned to treatment. The superiority of CEA was apparent. The 30-day incidence of any stroke or death was significantly higher for CAS compared with CEA (9.6%; 95% CI 6.4–14.0%; vs 3.9%; 95% CI 2.0–7.2%, respectively) with a relative risk (RR) of 2.5 (95% CI 1.2–5.1). 10 The absolute risk increase of CAS was 5.7%, suggesting that one additional stroke or death resulted when 17 patients underwent CAS instead of CEA. 10 Similar results were demonstrated for the 30-day incidence of disabling stroke or death (3.4%; 95% CI 1.7–6.7; vs 1.5%; 95% CI 0.5–4.2, for CAS vs CEA, respectively; RR 2.2%; 95% CI 0.7–7.2). No difference was seen between the centers when the experience of the performing physician (experienced, tutored during training, or tutored after training) or the hospital workload (< 21, 21–40, or > 40 patients/center enrolled) were taken into account. 10
EVA-3S recently reported its late results. 11 At the 4-year follow-up, 29 of the 265 patients who were assigned to CAS and 15 of the 262 patients who were assigned to CEA had had a stroke or died within 30 days of the procedure or had a nonprocedural ipsilateral stroke (cumulative probability: 11.1% vs 6.2%, respectively; HR 1.97; 95% CI 1.06–3.67; p = .03). 11 For any stroke or periprocedural death, the HR was 1.77 (95% CI 1.03–3.02; p = .04), whereas for any stroke or death, the HR was 1.39 (95% CI 0.96–2.00; p = .08). 11 A hazard function analysis showed that despite the fact that the 4-year estimated cumulative risks of periprocedural stroke or death and nonprocedural ipsilateral stroke were significantly higher after CAS than after CEA, this difference was largely accounted for by the higher periprocedural risk of CAS compared with CEA. Thus, the risk of ipsilateral stroke beyond the perioperative period was low and did not vary significantly between the two forms of treatment. 11 The conclusion reached was that although these results suggest that CAS is as effective as CEA for medium-term prevention of ipsilateral stroke, at least for the first 4 years after the perioperative period, “the safety of CAS needs to be improved before it can be used as an alternative to CEA in patients with symptomatic carotid artery stenosis.” 11
Stent-Protected Angioplasty versus Carotid Endarterectomy (SPACE)
In SPACE, 1,214 patients with symptomatic carotid artery stenosis were randomly assigned within 180 days of a TIA or a stroke episode either to CAS (n = 613 patients) or to CEA (n = 601 patients). 12 The primary end point was ipsilateral ischemic stroke or death from the time of randomization up until 30 days following the procedure. 12 SPACE failed to prove the noninferiority of CAS compared with CEA for rates of ipsilateral stroke or death within 30 days. The rate of death or ipsilateral ischemic stroke from randomization to 30 days after the procedure was 6.84% for CAS and 6.34% for CEA (absolute difference 0.51%; 95% CI −1.89 to 2.91; p = .09). 12
At the 2-year follow-up report of SPACE, the superiority of CEA over CAS with regard to restenosis (as defined ultrasonographically) was apparent. 13 A total of 541 (89%) of the patients in the CAS group and 522 (89%) of the initial study participants in the CEA group in the intention-to-treat population had clinical follow-up data at 2 years. 13 There were no significant differences in clinical secondary end points at 2 years between patients undergoing CAS compared with those undergoing CEA. Nevertheless, a ≥ 70% recurrent carotid stenosis (as measured by ultrasound criteria) was significantly more common in the CAS compared with the CEA group in both the intention-to-treat (10.7% vs 4.6%, respectively; p = .0009) and per-protocol (11.1% vs 4.6%, respectively; p = .0007) life table estimates. 13 Most restenoses occurred in the first 6 months after treatment (28 vs 12 restenoses in the CAS vs CEA groups, respectively). The clinical significance of recurrent stenosis after CAS remains unclear. It may be argued that this restenosis may be of little clinical consequence. An opposite opinion, however, could be that although only two instances of recurrent stenoses after CAS led to the development of neurologic symptoms, 13 this number might increase with a longer follow-up period. Within SPACE, the subgroup analyses revealed an interesting finding: there were differences in the 2-year risk of ipsilateral stroke between CAS and CEA. 13 The excess risk associated with stenting was greater in men than in women (p = .03), in patients aged ≥ 70 years (p = .08), and in those with a previous stroke (p = .12). 13 Therefore, CAS may provide equal or even better results than CEA in selected patient groups (eg, in females < 70 years of age).
Discussion
Our literature search revealed one single-center 7 and three multicenter randomized studies 8–13 comparing the long-term results of CEA with CAS in symptomatic patients (see Table 1). All four studies independently reached the conclusion that CAS provides inferior long-term results compared with CEA for the management of symptomatic carotid artery stenosis. 7–13 A higher rate of recurrent carotid stenosis 7,9,13 and a higher incidence of stroke or periprocedural death 11 are the main drawbacks for CAS compared with CEA. These results indicate that CAS may not provide equivalent long-term results to CEA for the management of symptomatic carotid artery stenosis.
Positive results for the supporters of CAS had been reported in two multicenter trials comparing CEA with CAS, namely, the Carotid Revascularization Using Endarterectomy of Stenting Systems (CaRESS) study 16 and the Stenting and Angioplasty with Protection in Patients at High Risk for Endarterectomy (SAPPHIRE) trial. 17,18 In CaRESS, only 33% of the CEA patients and 31% of the CAS individuals were symptomatic. 16 Similarly, most patients included in SAPPHIRE (71%) had asymptomatic stenosis, which carries a lower risk of stroke during carotid repair compared with symptomatic carotid stenosis. 17–19 Furthermore, CaRESS was not a randomized trial, 16 and in SAPPHIRE, less than half of the patients were randomized between CAS and CEA. 20
CAS versus CEA: Results from Systematic Reviews and Meta-Analyses
Although a different opinion is held by some, 21,22 the results of most systematic reviews and meta-analyses lend support to the superiority of CEA over CAS. 23–32 Among these, the four latest meta-analyses are considered as these included the most recent studies. The first meta-analysis (10 trials; 3,580 patients) showed that when a subgroup analysis of trials was performed including only symptomatic patients with a ≥ 60% carotid stenosis, the 30-day risks for both any stroke or death (RR 1.63; 95% CI 1.18–2.25 after CAS use; p < .05) and stroke (RR 1.62; 95% CI 1.13–2.31 after CAS use; p < .05) were significantly higher in CAS patients, “making it a suboptimal choice for symptomatic patients with moderate to severe stenosis.” 23 As the authors themselves admitted, however, “many of the trials included in this meta-analysis were not performed on patients with high surgical risk.” As a result, it could not be concluded “whether CAS or CEA is better for the truly high-risk surgical patient.” 23
In the second meta-analysis (8 trials; 2,985 patients with 2,646 [89%] symptomatic patients), a significant increase of 38% was demonstrated in the odds of suffering any stroke or death within 30 days after endovascular treatment compared with CEA (odds ratio [OR] 1.38; 95% CI 1.04–1.83; p = .024). 24 However, a significant heterogeneity was seen (Cochran Q = 14.0; p = .03) with an I 2 of 57.1%. After exclusion of non–peer-reviewed and small trials, as well as those trials with a majority of asymptomatic patients, the combined OR was smaller and nonsignificant (OR 1.29; 95% CI 0.94–1.76; p = .11). Although no significant heterogeneity was found in this analysis (Cochran Q = 4.5; p = .10), an I 2 value of 55.9% was found, which can still be regarded as substantial heterogeneity. Similarly, in this meta-analysis, after combining all trials, a nonsignificant increase in the odds of suffering a disabling stroke or death was demonstrated in the CAS compared with the CEA groups (OR 1.37; 95% CI 0.92–2.04; p = .12). In this analysis, again, no significant heterogeneity was seen (Cochran Q = 5.2; p = .27) with a substantially lower value of I 2 (22.8%). Finally, in the analysis of the large trials with symptomatic patients, a nonsignificant increase in the odds of suffering a disabling stroke or death was demonstrated for CAS compared with CEA (OR 1.33; 95% CI 0.89–1.93; p = .17). As before, no significant heterogeneity was found (Cochran Q = 1.1, p = .58) with an I 2 value of 0%. Based on these results, the conclusion reached was that “surgical treatment is still the gold standard for treatment of patients with symptomatic carotid artery stenosis, who do not have an increased surgical risk…. CAS has not yet been shown to be as safe or even safer than CEA in large clinical trials when considering short-term stroke and death rates.” 24
The third meta-analysis (10 randomized controlled trials with 3,182 patients) 25 showed that at 30 days and compared with CEA, CAS was associated with a nonsignificant reduction in the risk of death in 5 studies (RR 0.61; 95% CI 0.27–1.37; I 2 = 0%) but a nonsignificant increase in the risk of any stroke (RR 1.29; 95% CI 0.73–2.26; I 2 = 40%). When only major and disabling strokes were included in the analysis, a similar nonsignificant increase in the risk of stroke was noted in patients who underwent CAS in four studies (RR 1.06; 95% CI 0.32–3.52; I 2 = 45%). The conclusion reached was that although the impact on stroke remains unestablished from the results of this meta-analysis, “these results are consistent with a clinically important increase in stroke risk with CAS, an intervention that aims at reducing the risk of stroke.” 25
The fourth and most recent meta-analysis (8 studies; 2,942 both symptomatic and asymptomatic patients; 1,462 CEA and 1,480 CAS procedures) did not show any difference between CEA and CAS patients in terms of the incidence of stroke and death within 30 days or 1 year after the procedure, but the data favored CEA (at 30 days: 87 of 1,462 vs 121 of 1,480 events for CEA vs CAS, respectively; RR 0.69; 95% CI 0.45–1.07; p = .10; at 1 year: 68 of 511 vs 69 of 522 events for CEA vs CAS, respectively; RR 0.88; 95% CI 0.43–1.79; p = .72). 26 However, when considering only previously symptomatic patients, CAS was associated with a higher incidence of stroke and death within 30 days (54 of 1,011 vs 90 of 1,032 events for CEA vs CAS, respectively; RR 0.53; 95% CI 0.30–0.95; p = .003). 26 Furthermore, CAS had higher 1-year restenosis rates (7 of 253 vs 25 of 251 for CEA vs CAS, respectively; RR 0.28; 95% CI 0.28–0.63; p < .01). 26
In the report of the Vascular Registry of the Society for Vascular Surgery Outcomes Committee, 6,403 procedures with discharge data were entered by 287 providers from 6 specialties at 56 centers from July 11, 2005, to December 26, 2007. 27 The 30-day combined postprocedural risk of death, stroke, and myocardial infarction was higher for CAS compared with CEA in both symptomatic (46 of 645 vs 19 of 506, or 7.13% vs 3.75%, for CAS vs CEA, respectively) and asymptomatic patients (37 of 805 vs 17 of 862, or 4.60% vs 1.97%, for CAS vs CEA, respectively). 27 Although patients undergoing CAS had a higher incidence of preprocedural lateralizing neurologic symptoms (preprocedural stroke or TIA episode; 49.2% vs 42.4% for CAS vs CEA, respectively; p < .001), better outcomes were still demonstrated for CEA compared with CAS after risk adjustment for age, history of stroke, diabetes, and American Society of Anesthesiologists grade (ie, factors found to be significant confounders in outcomes using backwards elimination) by logistic regression analysis. 27
Finally, a nationwide survey in the United States of all carotid revascularization procedures performed in 2005 (135,701 CEA and CAS procedures; CEAs: 91%, CAS procedures: 9% ) showed that, compared with CEA, CAS was associated with both increased postoperative stroke (1.1% vs 1.8%, respectively; OR 1.7; 95% CI 1.2–2.3; p < .05), as well as overall mortality rates (0.57% vs 1.1%, respectively; OR 1.5; 95% CI 0.96–2.5; p < .05). 28 More importantly, the mortality difference increased considerably in patients with symptomatic disease (4.6% vs 1.4%, for CAS vs CEA, respectively; p < .05). The conclusion reached in this report was that CAS may not be an equal alternative to CEA for the management of patients with carotid artery stenosis. 28
Conclusions
Based on all of these results from recent randomized controlled trials and meta-analyses, it would seem that CAS has outcomes that are inferior to those of CEA for the treatment of carotid stenosis in patients with neurologic symptoms. However, it is premature to consider this conclusion definitive for several reasons. CAS technology and the technical expertise of operators currently performing the procedure are improving and are superior to those in the studies thus far reported. In addition, information is accumulating on how to use CAS more selectively and in which patients the procedure should be avoided. Elderly patients and those with an adverse aortic arch, supra-aortic branch, and cervical carotid anatomy are two clear examples. All of these developments may well improve the outcomes of CAS in symptomatic patients. Nevertheless, these improvements must be shown to do so by appropriately designed randomized controlled trials. Until such evidence exists, CEA should continue to be considered the primary treatment option for symptomatic carotid artery stenosis unless there are anatomic or hostile neck considerations such as a high lesion or cervical scarring or infection. 23–32
However, the jury regarding the role of CAS is still out, and its place in the treatment of symptomatic carotid stenosis could easily change as more evidence accumulates. Such evidence may come from two multicenter randomized trials, the Carotid Revascularization Endarterectomy vs Stenting Trial (CREST) and the International Carotid Stenting Study (ICSS). Both of these trials will report their results in 2009, and they will certainly have a considerable impact on determining the role of CAS in the management of patients with carotid stenosis. However, it is possible that they, too, will not be definitive, and the final judgment regarding CAS in symptomatic patients will come only after additional trials, some of which may have to be performed in groups of selected patients treated by experienced and skilled operators using improved optimal carotid stent technology and techniques.
Footnotes
Acknowledgment
Financial disclosure of authors and reviewers: None reported.