Author + information
- Received July 23, 2007
- Revision received September 4, 2007
- Accepted September 17, 2007
- Published online January 29, 2008.
- Sriram S. Iyer, MD, FACC⁎,⁎ (, )
- Christopher J. White, MD, FACC†,
- L. Nelson Hopkins, MD‡,
- Barry T. Katzen, MD§,
- Robert Safian, MD, FACC∥,
- Mark H. Wholey, MD¶,
- William A. Gray, MD, FACC#,
- Rocco Ciocca, MD⁎⁎,
- William B. Bachinsky, MD, FACC††,
- Gary Ansel, MD‡‡,
- James D. Joye, DO, FACC§§,
- Mary E. Russell, MD, FACC⁎⁎,
- BEACH Investigators
- ↵⁎Reprint requests and correspondence:
Dr. Sriram Iyer, Lenox Hill Hospital, 130 East 77th Street, Black Hall, 9th Floor, New York, New York 10021.
Objectives The multicenter, single-arm BEACH (Boston Scientific EPI: A Carotid Stenting Trial for High-Risk Surgical Patients) evaluated outcomes in high-surgical-risk patients with carotid artery stenosis treated with the Carotid WALLSTENT plus FilterWire EX/EZ Emboli Protection System (Boston Scientific, Natick, Massachusetts).
Background Carotid artery stent (CAS) placement offers a less invasive alternative for high-risk surgical carotid endarterectomy (CEA) patients.
Methods The trial enrolled 480 pivotal patients who were candidates for carotid revascularization but considered high surgical risk due to pre-specified anatomic criteria and/or medical comorbidities. The primary end point (all stroke, death, or Q-wave myocardial infarction [MI] through 30 days; non–Q-wave MI through 24 h; and ipsilateral stroke or neurologic death through 1 year) was compared with a proportionally weighted objective performance criterion (OPC) of 12.6% for published surgical endarterectomy results in similar patients, plus a pre-specified noninferiority margin of 4%.
Results Among pivotal patients, 41.2% were at high surgical risk due to comorbid risk factors, and 58.8% due to anatomic risk factors; 76.7% were asymptomatic with flow-limiting carotid stenosis >80%. At 1 year, the composite primary end point occurred in 8.9% (40 of 447), with a repeat revascularization rate of 4.7%. With an upper 95% confidence limit of 11.5% for the primary composite end point, the BEACH trial results met the pre-specified criteria for noninferiority relative to the calculated OPC plus noninferiority margin (16.6%) for historical surgical CEA outcomes in similar patients (p < 0.0001 for noninferiority).
Conclusions The BEACH trial results demonstrate that CAS with the WALLSTENT plus FilterWire embolic protection is non-inferior (equivalent or better than) to CEA at 1-year in high-surgical-risk patients (Boston Scientific Embolic Protection, Inc. [EPI]: A Carotid Stenting Trial for High-Risk Surgical Patients [BEACH]; http://clinicaltrials.gov/ct2/show/NCT00316108?term=NCT00316108&rank=1; NCT00316108).
Carotid artery disease is a major cause of ischemic stroke, with an absolute risk directly related to the severity of stenosis and presence of neurologic symptoms (1,2). Surgical removal of atherosclerotic plaque (carotid endarterectomy [CEA]) can successfully reduce the incidence of cerebral infarction in both symptomatic patients and asymptomatic patients with carotid stenosis >60% (3–7). Some patients, however, are unable to safely undergo CEA due to unfavorable anatomy or comorbid conditions (8–11), and are increasingly considered for treatment by carotid artery stent (CAS) placement with cerebral embolic protection (12–19). The BEACH (Boston Scientific EPI: A Carotid Stenting Trial for High-Risk Surgical Patients) trial is an ongoing, prospective, multicenter, single-arm trial evaluating the outcomes of high-surgical-risk patients treated with the Carotid WALLSTENT and either the FilterWire EX or FilterWire EZ Emboli Protection System distal filter device (Boston Scientific Corp., Natick, Massachusetts). Thirty-day outcomes for the pivotal group were reported previously (20); this report discusses 1-year pivotal outcomes.
The BEACH trial was designed to test if outcomes in high-surgical-risk patients treated with the WALLSTENT plus FilterWire EX/EZ would be non-inferior to a calculated objective performance criterion (OPC) agreed upon with the U.S. Food and Drug Administration (FDA). The primary end point was 1-year composite morbidity and mortality. Noninferiority was assessed by comparing the upper 95% confidence interval (CI) of the observed end point to a literature-derived OPC for CEA in patients with similar demographics, with a pre-specified noninferiority margin (delta) of 4%.
Patients were enrolled into a roll-in phase to familiarize physicians with the protocol and devices (first 1 to 9 patients per site), a pivotal phase, or a bilateral registry (patients requiring treatment for both carotid arteries). After review of roll-in periprocedural outcomes, investigators were allowed to proceed to the pivotal phase. This report discusses 1-year outcomes, but study patients continue to be monitored through 3 years of follow-up.
Patient selection, device, procedure, and follow-up
Details of patient selection, device, procedure, and follow-up have been described (20). Briefly, eligible patients with carotid disease met general inclusion criteria plus at least 1 definition of surgical high risk based on specific anatomic and comorbid clinical criteria (Table 1). The common carotid artery (CCA), bifurcation, or internal carotid artery (ICA) was 4 to 9 mm in diameter, with ≥50% stenosis by angiography in symptomatic patients and ≥80% in asymptomatic patients as determined by the operator (visual estimate) per NASCET (North American Symptomatic Carotid Endarterectomy Trial) criteria (7). The protocol was approved by institutional review boards of the 47 participating U.S. institutions; written informed consent was obtained from patients before enrollment, and the trial complied with the Declaration of Helsinki. Trial results are reported on the National Institutes of Health website.
Election of stent size (Carotid WALLSTENT Monorail Endoprosthesis, Boston Scientific Corp.) and emboli protection length and placement (FilterWire EX or FilterWire EZ Embolic Protection System, Boston Scientific Corp.) were based on the operator’s visual estimate of vessel diameter. At the end of the procedure, the filter together with any contained material was collapsed and retrieved. All patients were examined before (within 7 days) and after the procedure (per protocol, within 24 h and at the time of any change in clinical symptoms) by an independent neurologist or neurosurgeon certified in the administration of the National Institutes of Health Stroke Scale (NIHSS). Patients were also monitored continuously during the procedure and frequently during the in-hospital recovery period by physician and nursing staff, including neurologic examinations. All patients underwent carotid duplex ultrasonography before the procedure and before discharge. Investigator training included review of animal studies, on-site proctoring to achieve competence in device implantation, and general training on the protocol. Independent ultrasound and angiographic core laboratories provided review of all studies throughout the course of the trial and validation of site-determined entry criteria. Follow-up included carotid duplex ultrasonography as well as independent neurologic examination using the NIHSS at 1, 6, and 12 months and yearly thereafter through 3 years.
Primary end point
The primary composite end point included all stroke, death, and Q-wave myocardial infarction (MI) through 30 days; non–Q-wave MI through 24 h; and ipsilateral stroke and neurologic death through 1 year. Adverse events were adjudicated by an independent clinical events committee. Analyses of all measurements obtained were performed according to published criteria by core laboratories (Online Appendix 1) (20).
Harvard Clinical Research Institute (Online Appendix 1) performed data management and statistical analyses with SAS version 8.2 or above (SAS Institute Inc., Cary, North Carolina). Freedom from 1-year morbidity and mortality was determined by Kaplan-Meier analysis; predictors were identified using single- and multi-variable logistic regression analyses (p < 0.05 for significance).
In the BEACH trial, CAS was compared to CEA (control) to determine if stenting was non-inferior to surgery in the high-surgical-risk patient population. The CEA comparator was a calculated OPC applied to the primary composite end point. The mathematical equation (in the following text) representing the OPC was derived using data in the literature (Online Appendix 2) from CEA outcomes in high-surgical-risk patients and reflects the calculated surgical risk associated with anatomical conditions (πOPC - Anatomical) plus that associated with comorbid conditions (πOPC - Comorbid). The weighted OPC (πOPC - Weighted), reflecting the calculated CEA risk for the BEACH primary end point, was obtained by multiplying the anatomical and comorbid risk by the percentage of BEACH patients in the surgical high risk anatomical (ωA) and surgical high risk comorbid (ωC) groups, respectively, as shown in the equation below.
With the concurrence of the U.S. FDA, 11% was selected as πOPC - Anatomical and 15% as πOPC - Comorbid, with the higher risk assumed in cases qualifying for both. Using the enrollment percentages in each category (58.8% anatomical [ωA] and 41.2% comorbid [ωC]), the calculated weighted OPC for this trial was 12.6%. Because true equivalency cannot be realized with less than an infinite sample size, an FDA-concurred pre-specified spread of 4% for the “delta” definition of equivalency was added to the weighted OPC to yield an upper noninferiority boundary of 16.6%. The BEACH primary composite end point with a 1-sided upper 95% CI was compared with this boundary (weighted OPC plus delta) to assess noninferiority of CAS to CEA. Statistical significance was determined using a normal approximation test for a single proportion (Z test).
The BEACH trial enrollment began February 2002 and ended December 2003 with 189 roll-in phase patients, 78 patients in the bilateral registry, and 480 patients in the pivotal study group. The trial was temporarily suspended on December 24, 2002 due to carotid WALLSTENT malfunctions and re-initiated on June 11, 2003 after the cause of the malfunctions was identified and corrected. Of 47 centers participating (Online Appendix 3), 36 contributed to the pivotal group. Data for patients receiving the FilterWire EX and FilterWire EZ were pooled because there was no statistically significant difference in end points between these groups after adjustment for baseline percent diameter stenosis.
Baseline and procedural characteristics—pivotal group
Among pivotal patients, 41.2% had comorbid risk factors, and 58.8% had anatomical risk factors, resulting in a calculated noninferiority margin (OPC plus 4% delta) of 16.6% for the 1-year composite primary end point. Table 2 lists patient qualifications by individual high-risk criteria; Table 3 lists baseline patient and lesion characteristics. Over three-quarters of the patients (77%, 368 of 480) were asymptomatic; 99.2% of whom had site-reported (visual estimate) carotid stenosis between 80.0% and 99.9%. Quantitative analysis of percent diameter stenosis determined by the core lab for the entire pivotal group was 71.6 ± 10.7% (mean ± standard deviation); this decreased to 10.6 ± 14.4% post-procedure (Fig. 1) with 99.8% of patients having <50% post-procedure stenosis. Procedure success—a composite end point based on patients in whom a system placement attempt was made and including system technical success (successful delivery, deployment, and retrieval of the devices; 98.3%); angiographic success (in-stent residual diameter stenosis ≤30% post-procedure; 90.8%); and the absence of death, stroke, or MI within 24 h of the index procedure—was 87.6% (418 of 477).
Clinical outcomes—pivotal group
At 30 days the composite major adverse event rate of death, stroke, and Q-wave MI in the pivotal group was 5.4% (24 of 447) (Table 4). The 1-year primary composite end point was 8.9% (40 of 447 patients) (Table 4). The 1-sided 95% upper CI was 2.6%, resulting in an upper limit (central estimate plus 95% CI) of 11.5% for the BEACH primary end point. This value was below the control comparator of 16.6% (calculated OPC for similar CEA patients of 12.6% plus the pre-specified delta of 4%), thereby indicating noninferiority of CAS to CEA (p < 0.0001 for noninferiority).
Table 5 shows primary end point outcomes for asymptomatic and symptomatic patients. There were no significant differences in rates between the 2 groups though there was a weak trend towards a higher rate of the primary end point for symptomatic versus asymptomatic patients at 1 year (12.5% vs. 7.8%, respectively, p = 0.14). Patients qualifying for the trial with comorbid high-surgical-risk factors had a significantly higher 30-day event rate than did those with anatomical high-risk factors (14.3% vs. 5.3%, respectively, p = 0.002). Advanced age (≥75 years) was associated with significantly worse outcomes, including late neurologic death (Fig. 2).
Freedom from morbidity and mortality was 91.6% at 1 year with most events occurring at ≤30 days (Fig. 3). Outcomes from 31 to 360 days (Table 6) include a late neurologic event rate related to stenting of 2.7% (12 of 447). The 1-year restenosis rate, defined as ≥70% stenosis by duplex ultrasound, was 8.9% (40 of 447), with a 1-year repeat target vessel revascularization rate of 4.7% (20 of 425) including 17 asymptomatic patients with repeat revascularization driven by duplex findings alone. Per the BEACH trial protocol, a carotid angiogram was required if the duplex ultrasound demonstrated a restenosis of ≥70%. This potentially increased the re-intervention rate above what would be expected for this primarily asymptomatic group.
Predictors of 1-year morbidity and mortality—pivotal group
By univariate analysis, age ≥75 years (p = 0.0004) and the comorbid risk category (p = 0.0017) were significant predictors of 1-year morbidity and mortality. Multivariate analysis identified these same significant predictors (p = 0.0002 and p = 0.0092, respectively) plus diabetes (p = 0.0357) and symptomatic status (p = 0.0774).
Durability of revascularization—ultrasound results in the pivotal group
The ICA/CCA peak systolic velocity (PSV) ratio (mean ± standard deviation) before the procedure was 5.3 ± 3.1. This improved to 1.4 ± 0.5 immediately after the procedure and remained improved at 6 months (1.9 ± 1.2) and 1 year (1.9 ± 1.1). Additionally, there was no progression of ICA PSVmaximum over the latter 6 months of follow-up (Table 7).
This prospective, multicenter, single-arm trial compared the outcomes of carotid stenting with a distal emboli protection filter to a calculated OPC based on the historical rate of similar end points in high-surgical-risk patients undergoing CEA. The observed composite 1-year morbidity and mortality rate was 8.9%, with a 95% 1-sided upper confidence limit extending to 11.5%. This was significantly (p < 0.0001) below the 16.6% noninferiority boundary defined as the 12.6% historical rate in comparable surgical patients plus the pre-specified noninferiority margin of 4%. Therefore, the BEACH trial demonstrates that this form of CAS with distal embolic protection using the WALLSTENT in combination with the FilterWire emboli protection system is noninferior to (equivalent or better than) surgical CEA in high-surgical-risk patients (18,19).
Moreover, most of the morbidity and mortality in the 1-year primary end point was due to early events. The rate for death, stroke, and Q-wave MI was 5.4% (CI 3.5% to 7.9%) through 30 days; most events (3.4% [CI 2.8% to 6.8%]) were ipsilateral stroke (major [1.1%], minor [2.0%], hemorrhagic [0.2%]) (20). This is similar to the 4.8% rate of death, stroke, or MI in the randomized SAPPHIRE (Stenting and Angioplasty with Protection in Patients at High-Risk for Endarterectomy) trial, which included 29.9% symptomatic patients in the stenting arm (18). The rate for symptomatic BEACH patients (7.7%) is similar to the 30-day event rate in 2 recent prospective randomized clinical trials of symptomatic carotid occlusive disease patients that failed to show a benefit of stenting over surgery (21,22). In the SPACE (Stent-Supported Percutaneous Angioplasty of the Carotid Artery versus Endarterectomy) trial, the 30-day 6.8% rate of death or ipsilateral ischemic stroke in the stenting arm was not statistically different from the 6.3% rate in the CEA arm, and failure to demonstrate noninferiority of CAS to CEA despite similar event rates reflects premature termination of enrollment at 1,183 patients, resulting in inadequate statistical power (21). The EVA-3S (Endarterectomy versus Angioplasty in Patients with Symptomatic Severe Carotid Stenosis) trial reported a 9.6% incidence of any stroke or death in the stenting arm at 30 days with a corresponding CEA rate of 3.9%, possibly reflecting relative inexperience of CAS operators compared with very experienced surgical operators, as well as premature termination of the trial for safety concerns (22). In addition, comparison between U.S. and European studies may be limited as there is no distinction between patients considered high-risk or non–high-risk for surgery outside the U.S. The results of these 2 symptomatic trials, thus, are outliers from the lower CAS rates seen in most of the high-surgical-risk carotid stenting trials.
In the BEACH trial, carotid duplex ultrasound at 6 months and 1 year indicated continued vessel patency, as reflected in a 1-year restenosis rate of 8.9%, a target vessel revascularization rate of 4.7%, and a stent-associated late neurologic event rate of 2.7%. The overall 8.6% composite 1-year end point in the BEACH trial also compares favorably to earlier trials. In the SAPPHIRE trial, the primary end point (death, stroke, or MI at 30 days plus ipsilateral stroke or death from neurologic causes within 31 days to 1 year) was 12.2% (18). In the high-risk stent registry ARCHeR (Acculink for Revascularization of Carotids in High Risk Patients) (23.8% symptomatic patients), the primary end point (death, stroke, MI at 30 days plus ipsilateral stroke at 1 year) was 9.6% (19). In this high-risk patient population, confounding coronary and pulmonary morbidities are common, but their influence is limited by the use of a 1-year end point restricted to neurologic death and ipsilateral stroke beyond 30 days. The concordance in pure neurologic outcomes for the BEACH trial and 1-year results reported in 2 other large studies involving carotid stenting and high-surgical-risk patients with similar end points indicates that CAS can be reproducibly achieved in this patient population.
Older age was identified as a significant predictor of poor outcome as patients 75 years or older experienced higher event rates. Increased procedural risk may be attributable, in part, to a number of factors commonly associated with advanced age including excessive vessel tortuosity, arch elongation, and heavy calcification, which also contribute to increased risk with stent placement (23). Plaque burden is presumed to increase with age, leading to more events in the elderly population independent of intervention (24). In the lead-in phase of the CREST (Carotid Revascularization Endarterectomy vs. Stent Trial) (30.7% symptomatic), octogenarians exhibited increased complications at 30 days with CAS (25). Higher risk has been seen in elderly patients for both stenting and surgery, although improved devices and experience in CAS along with better patient selection have led to reductions in perioperative morbidity (26–33). Advances notwithstanding, the observed CAS results in older patients highlight the need to use CAS with distal emboli protection predominantly in patients at low risk for stenting, unless their risks for CAS are elevated to an even greater degree. Trials that enroll a high proportion of high stent risk patients would similarly be expected to show higher complication rates, and operators should remember that medical management, in fact, may be a better choice for patients deemed high risk for both CAS and CEA (34).
It is critical to recognize that there has been a tremendous amount of progress made since the BEACH trial was initiated in 2002 regarding patient selection as a determinant of CAS outcomes. While the BEACH trial was designed to evaluate patients considered high risk for endarterectomy, it did not exclude those patients who would be considered high risk for stenting; the BEACH trial met the primary end point with a major adverse event rate below the FDA-agreed OPC. In asymptomatic patients with severe carotid stenosis, the 30-day event rate was 4.7%, which compares favorably with the asymptomatic SAPPHIRE group (5.4%) (18). However, recently drafted multispecialty consensus statements have concluded that with improved understanding of appropriate patient selection, the acceptable 30-day event rates for carotid stenting should be ≤3% for asymptomatic patients and ≤6% for symptomatic patients (35). The adoption of CAS as a minimally invasive alternative to CEA will be dependent upon both proper patient selection and appropriate physician training and experience. Further clinical trials should, thus, evaluate patients who are considered low or normal risk for both CEA and CAS—the majority of patients with carotid artery occlusive disease. Only positive results from this type of trial will move CAS from a niche procedure to a mainstream treatment option or the dominant standard of care.
The 1-year primary end point of 8.9% (upper confidence limit of 11.5%) for carotid stenting and emboli protection in high-surgical-risk patients treated in the BEACH trial meets the criteria for noninferiority to the 16.6% severity-adjusted surgical OPC plus delta. In high-surgical-risk patients who meet indications for carotid revascularization, CAS with emboli protection is not inferior to CEA at 1 year.
The authors thank Alex Shih, BS, Lynne Kelley, MD, and Ruth Starzyk, PhD (Boston Scientific Corp.), for their assistance in the statistical analyses (A.S.) and in drafting this manuscript (L.K. and R.S.).
For a list of core laboratories (Online Appendix 1), references used in the OPC calculation (Online Appendix 2), and investigators and institutions participating in the BEACH trial (Online Appendix 3), please see the online version of this article.
The BEACH trial was supported by Boston Scientific Corporation. The current affiliation for Dr. Ciocca is University of Massachusetts Medical School, Worcester, Massachusetts. The current affiliation for Dr. Russell is Ascent Translational Sciences, Carlisle, Massachusetts.
- Abbreviations And Acronyms
- carotid artery stent/stenting
- common carotid artery
- carotid endarterectomy
- confidence interval
- Food and Drug Administration
- internal carotid artery
- myocardial infarction
- National Institutes of Health Stroke Scale
- objective performance criterion
- peak systolic velocity
- Received July 23, 2007.
- Revision received September 4, 2007.
- Accepted September 17, 2007.
- American College of Cardiology Foundation
- Bond R.,
- Rerkasem K.,
- Rothwell P.M.
- Shawl F.,
- Kadro W.,
- Domanski M.,
- et al.
- Safian R.D.,
- Bresnahan J.F.,
- Jaff M.R.,
- et al.,
- CREATE Pivotal Trial Investigators
- Roubin G.S.,
- Iyer S.,
- Halkin A.,
- Vitek J.,
- Brennan C.
- Störk S.,
- van den Beld A.W.,
- von Schacky C.,
- et al.
- Setacci C.,
- de Donato G.,
- Chisci E.,
- et al.
- Roubin G.S.,
- New G.,
- Iyer S.S.,
- et al.
- McCrory D.C.,
- Goldstein L.B.,
- Samsa G.P.,
- et al.
- Zahn R.,
- Ischinger T.,
- Hochadel M.,
- et al.,
- Arbeitsgemeinschaft Leitende Kardiologische Krankenhausärzte (ALKK)
- Bates E.R.,
- Babb J.D.,
- Casey D.E.,
- et al.