Filter cerebral/embolic protection devices, commonly referred to as umbrella devices, offer a new, innovative approach to reducing the risk of peri-procedural neuroembolism and its associated neurologic sequelae during carotid artery angioplasty and stenting (CAS) in patients with severe carotid artery occlusive disease (21). Over the years, performance of CAS has been limited by an unacceptably high combined incidence of stroke and death (>6 percent), attributed to the distal embolization of debris released from the carotid plaque along the arterial wall during balloon dilation and/or stent deployment (23). Filter protection devices work by capturing such debris, usually within collapsible, filter-like baskets, preventing it from traveling to the brain and causing embolic neurologic complications (27).
OBJECTIVE
This study provides an overview of treatment options for severe carotid artery disease (CAD) and critically assesses available evidence pertaining to the efficacy and cost of using filter protection devices during CAS in the management of high-risk patients with severe CAD.
BACKGROUND
Incidence and Burden of Stroke in Canada and Abroad
Over the past decade, several studies have highlighted the growing importance of stroke as a major public health concern and contributor to the global burden of disease (5). In 1998, an estimated 5.1 million individuals died of the condition, making it the second leading cause of mortality worldwide. Furthermore, approximately 15 million individuals suffered a non-fatal case, 20 percent of whom experienced a second event within 5 years (28). Given the ageing population, these numbers are expected to escalate over the next 15 years, with stroke becoming the fourth greatest cause of combined death and disability worldwide by 2020 (22).
In Canada, approximately 45,000 strokes are reported each year. As the fourth leading cause of death and leading cause of disability among the country's adult population, stroke accounts for a large portion of overall health care spending (e.g., continuing-care and rehabilitation programs) and costs the Canadian economy roughly $2.7 billion dollars annually (26). Its chronic effects also impose a significant emotional burden, which is difficult to express meaningfully in monetary terms, on family members and caregivers (21).
Causes and Grades of Severity
Approximately one quarter of stroke cases are attributable to CAD, a disease characterized by carotid stenosis or narrowing of the carotid artery (9). This narrowing occurs when plaque accumulates inside the arterial wall, decreasing its diameter and restricting blood flow. Over time, pieces can break off, travel toward the brain, become wedged in a smaller branch of the artery, and cause a stroke. Among patients with CAD, the risk of stroke depends on the degree of stenosis (21).
Severe (high-grade) stenosis is commonly defined as >70 percent vessel diameter reduction in symptomatic patients and >80 percent reduction in asymptomatic patients, as observed through either duplex ultrasonography or angiography (gold standard) (11). In general, symptomatic patients comprise those whose initial sign of disease is a transient ischemic attack (TIA; i.e., focal neurologic abnormalities, e.g., weakness or numbness on one side of the face or body, changes in vision, confusion, dizziness, double vision, slurred speech, and so on) of sudden onset and brief duration (less than 24 hours) caused by a temporary interruption of blood flow to a specific area of the brain) (18). Within 1 year of diagnosis, the risk of stroke among symptomatic and asymptomatic high-grade stenotic patients is approximately 13 percent and 5 percent, respectively (11).
Treatment
The accepted best-practice approach to treating such patients is carotid endarterectomy (CEA), involving surgical dissection of the artery and removal of the plaque from inside the vessel wall through a 3- to 4-inch incision along one side of the neck (20). CEA has been well validated in several large, multi-center, prospective, randomized, controlled trials, the findings of which clearly support its use over medical management (MM; administration of antiplatelet, anticoagulant, and lipid-lowering agents and risk-factor reduction) in high-grade CAD subpopulations (1-year stroke risks: MM, 12 percent [asymptomatic] and 22 percent [symptomatic] compared with CEA, 5 percent [asymptomatic] and 8 percent [symptomatic]) (1). However, despite its proven efficacy, operative risks exist (e.g., general anesthesia, hematoma, cranial nerve damage, infection, and so on), which may make surgery risky, if not impossible in some patients (31). Such high-risk patients typically manifest one or more co-morbidities (e.g., heart disease, pulmonary disease, or renal insufficiency), in addition to severe stenosis, and decisions regarding whether or not they can tolerate surgery are left to clinical judgment (11). It is believed that the majority of high-grade CAD patients managed in general clinical settings fall into this category, regardless of their symptom status. According to the results of a review of inclusion criteria for the trials designed to establish CEA's efficacy, over 75 percent of patients seen in daily practice would not have been eligible to participate (6). Nonetheless, options for treating high-risk patients may include CEA, MM, and/or CAS. Selection of the most appropriate approach depends upon whether or not even a small possibility of achieving a net benefit from CEA, after taking into account all operative risks, exists (Figure 1) (21).
Treatment algorithm for high-risk patients with severe carotid artery disease (CAD).
During CAS, a small puncture is made in the femoral artery, through which a catheter, wires, and a balloon are inserted and guided to the narrowed area of the carotid artery. There, the balloon is inflated to widen the artery and facilitate the delivery and placement of a stent (a small, stainless steel, expandable mesh tube). Once deployed, the stent becomes implanted into the arterial wall, acting as a scaffolding to keep the artery open. Several recent trials have demonstrated promising long-term outcomes and peri-procedural neurologic complication rates similar to that of CEA (11). However, the potential risk of stroke caused by plaque particles embolizing in the brain, either during or directly after the procedure, has been an obstacle to its widespread application (2). Advantages and disadvantages of all three treatment options are outlined in Table 1 (21).
Given growing recognition of patient preferences for safe, minimally invasive therapies that optimize health outcomes, a complete generation of dedicated cerebral (distal embolic) protection devices/systems (CPD) preventing embolic complications related to CAS has been developed (23). As mentioned above, these devices block the entry of plaque debris into the intracranial circulation by trapping them and allowing for their removal through a guiding catheter. Three classes of CPDs exist, each based upon a different approach to cerebral protection: (i) distal balloon occlusion, (ii) filter, or (iii) flow reversal (retrograde) (6). In distal balloon occlusion, a balloon-on-a-wircatheter containing an occlusion balloon is inflated in the internal carotid artery once it has crossed the target lesion or plaque. This strategy enables it to block the run-off circulation to the brain, trap any particulate debris generated during the stenting procedure, and, ultimately, create a standing column of blood, which can then be aspirated and discarded through an export catheter. Filter protection systems consist of an umbrella-like filter mounted on the end of a guidewire and delivery/retrieval catheters. The filter is advanced through the delivery catheter and positioned above the carotid lesion, where it is opened to form a tiny protective basket that catches freed plaque particles while maintaining blood flow. After placement of the stent, the basket is collapsed and pulled back through a retrieval catheter. In flow reversal systems, the direction of blood flow in the carotid artery is reversed, permitting capture and removal of plaque particles released during the procedure. To accomplish this reversal, balloons are first inflated to block the common and external carotid arteries. Then, a special type of arteriovenous shunt is used to reverse blood flow in the artery. A carotid-to-femoral arteriovenous shunt connecting a guiding catheter to a femoral vein introducer is then used to create reversal of flow in the carotid artery. This shunt contains an interposed blood filter to trap any debris generated during the procedure.
Early clinical experiences with prototype and first-generation devices in the mid to late 1990s have helped identify potential advantages and disadvantages of each cerebral protection approach (Table 2), and, in turn, led to refinements in the devices, themselves, as well as the steps required to achieve technical success (21). Despite this rapid evolution in all three classes, the filter principle has received the greatest level of clinical enthusiasm (6). This is, in part, because of its potential to trap particles of significant size while maintaining blood flow and its operational simplicity (relative to the other approaches).
Currently, several types of filter devices are manufactured, each of which contains a uniquely designed filter element. Although the market status, i.e., whether or not they have received regulatory approval for CAS, varies across countries, anecdotal evidence of “off-label” use of those licensed for other endovascular procedures (e.g., saphenous veins and coronary arteries) exists. This use has raised the need to examine the safety, efficacy, and economic implications of these devices (10).
METHODS
Literature Search
A comprehensive search for randomized or nonrandomized clinical trials, retrospective, prospective, or concurrent cohort studies, clinical series, or economic analyses comparing CEA or CAS without protection to CAS with cerebral protection and appearing in the international and national English-language literature between January 1990 and January 2005 was completed. Before conducting the search, critical terms describing stroke, carotid artery disease (stenosis), intracranial or cerebral emboli, stenting, angioplasty, neuroprotection, and cerebral protection were identified. These terms (free text and truncated), along with their likely synonyms, were first identified by examining the list of controlled vocabulary terms (e.g., medical subject heading terms [MeSH]) used to index known key references and then combined to form a structured search strategy reflecting important conceptual relationships between terms.
The following electronic bibliographic databases, which cover biomedical, clinical, social science, and health services management topics, were used to locate peer-reviewed research: PubMed/MEDLINE (Medical Literature Analysis and Retrieval System On-Line), EMBASE (Excerpta Medica Database), Cochrane Library, Biological Abstracts, HealthSTAR, CINAHL (Cumulative Index to Nursing and Allied Health Literature), EconLit, the Centre for Reviews and Dissemination databases (HTA [health technology assessment], NHS EED [National Health Service Economic Evaluation Database], and DARE [Database of Abstracts of Reviews of Effects]), Dissertation Abstracts International, and Web of Science (Science Citation Index, Social Science Citation Index, and Arts and Humanities Index). Unpublished or non–peer-reviewed information, i.e., gray literature; e.g., conference proceedings, technical papers, government sponsored reports, clinical newsletters, editorials, commentaries, and promotional material from manufacturers, was identified through Internet searches using meta-browsers capable of scanning multiple file formats (e.g., PDF files). In addition, a series of reference databases designed to capture grey literature was used. They include: SIGLE (System for Information on Grey Literature in Europe), Cabot (comprised of unpublished and published Canadian health services research literature), AMICUS (the National Library of Canada union catalogue of government and university library holding), the National Library of Medicine's Gateway, and NTIS (the United States National Technical Information Service). For completeness, the electronic search was supplemented by a manual search of (i) the most recent issues of vascular, neurological, radiological, and neurosurgical journals to identify articles not yet indexed in the bibliographic databases; and (ii) reference lists of papers selected for retrieval (Table 3).
Selection and Classification of Literature
Results of both the electronic and manual literature searches were imported into Reference Manager Version 10.0. After removal of duplicate entries, compiled citations (titles and abstracts, where available) were reviewed for relevancy by two independent reviewers using a predefined checklist composed of study inclusion criteria. Citations discussing CAS without mention of the use of cerebral protection devices, only non-filter cerebral protection systems, or angioplasty and stenting in vessels other than the carotid artery were excluded. Additionally, in cases where multiple citations appeared to report on the same patient population from a particular institution, all but the most recent one (determined by publication date and, when necessary, correspondence with the author) were removed. The remaining citations were subsequently retrieved. Selected literature was first sorted into groups according to methodological design and then further classified by publication status, i.e., whether they appeared in a published, peer-reviewed journal.
Synthesis and Critical Appraisal of Selected Literature
Information from studies was systematically extracted and recorded using a common data abstraction form, which contained key questions for assessing the purpose and methods (e.g., setting, number of patients, treatment protocol, outcome measures, and so on) of each study, as well as the validity of the results presented (8;12).
The quality of each study was scrutinized against a set of critical appraisal criteria, compiled from peer-reviewed, published guidelines previously designed to evaluate the strength and, in turn, validity, of evidence derived from specific types of clinical studies (non–randomized controlled trial [RCT] and cohort, RCT, or systematic reviews) was used, i.e., studies matched with guidelines based upon study design. The findings were then summarized in tabular form and analyzed to assess the overall efficacy of and economic implications associated with the use of cerebral protection devices during CAS.
RESULTS
Of the 560 discrete citations initially identified through the literature search, eight were selected for inclusion in the review: three clinical series, three cohort studies, one nonrandomized clinical trial, and one randomized clinical trial (Figure 2) (3;4;7;15;17;19;29;30). Seven of the studies reported events within 30 days of treatment, the period in which peri-procedural complications caused by the embolization of plaque particles (released during CAS) within the vessels of the brain are most likely to occur (3;7;15;17;19;29;30). The types and classes of cerebral protection devices used varied across studies, as did the characteristics of patients involved (e.g., elderly populations, high-risk populations, and so on), the scope of 30-day outcomes measured, and the strength of the evidence presented, which ranged from poor to good, when evaluated against accepted quality criteria (Table 4). Consequently, combining the data using meta-analytic techniques to calculate pooled incidence rates was deemed inappropriate.
Results of literature search and selection process.
Safety
Of the eight studies found, six provided information on technical or procedural success, i.e., the rate at which the CPD was successfully deployed and retrieved (3;4;7;15;17;29). Values ranged from 95.6 percent to 100 percent. Procedural complications reported, none of which had lasing effects, included transient spasm (contraction) of the artery in 2.9 percent to 7.9 percent of patients, blood flow impairment in 2.7 percent to 13 percent of patients, and arterial wall injury in 1 patient. Importantly, an acceptable risk of device-related procedural complications during CAS has been defined as that which does not exceed the rate of surgical complications observed with CEA, the gold standard treatment for CAD (19). The incidence of hematoma, cranial nerve palsies, cervical strain, and infection from CEA ranges from 5 percent to 27 percent (25). Some believe that the safety of filter protection devices is subject to the learning effect, i.e., failure rates decline as the number of procedures performed increases (14). To examine this, information on the points at which device deployment failures and major/minor complications occurred during a study or in clinical practice is required. None of the studies reviewed presented such information; therefore, it was not possible to determine the extent to which experience may have influenced the findings.
Efficacy of Cerebral Protection Devices
To establish the efficacy of CPDs, information evidence pertaining to their ability to reduce peri-procedural neurological complications caused by the embolization of plaque particles (released during CAS) within the vessels of the brain must exist (21). These complications, occurring within 30 days of the procedure, are transient ischemic attack (a focal retinal or hemispheric event from which the patient completely recovers within 24 hours), minor stroke (a new neurological deficit that either resolves completely within 30 days or increases the National Institutes of Health (NIH) stroke scale by 3), major stroke (a new neurological deficit that persists for 30 days and increases the NIH stroke scale by 4), and death (11). Thus, accepted primary clinical outcome measures/end points upon which to establish device efficacy include: 30-day risks of TIA, minor stroke, major stroke, or death, i.e., peri-procedural complications (6). For these devices to be proven efficacious, the number and severity of peri-procedural complications, assessed in similar groups of CAS patients treated with and without filter protection, must be less in the protected group (24). Furthermore, these rates should be equivalent to, if not lower than, that observed during CEA in the same types of patients, i.e., high-risk, high-grade CAD patients (13).
In three of the four studies comparing groups of patients who received CAS with cerebral protection to those who received only CAS, the 30-day incidence of death and stroke (major or minor) was statistically significantly greater in the latter group (7;17;19;30). Whereas the fourth study, a large clinical series, also reported a higher 30-day combined stroke/death incidence in patients treated with unprotected CAS, the difference did not achieve statistical significance (19).
None of the three studies comparing CAS with cerebral protection to CEA demonstrated a statistically significant difference in the 30 incidence of death, major stroke, or myocardial infarction between treatment groups (3;15;19). However, in the one study (a non-concurrent cohort study) that also reported the 30-day incidence of minor stroke, the rate was significantly higher among patients who received protected CAS (15).
Economic Implications
Despite an exhaustive search for economic analyses of cerebral protection devices, none could be located. Therefore, an attempt was made to calculate a crude, conservative estimate of potential economic implications, from a health care system perspective, under four assumptions: (i) Efficacy of filter protection devices. The 30-day risk of stroke and death for CEA is equal to that for CAS with filter protection; (ii) Indications for CAS with protection. Inoperable, high-grade CAD patients are all offered CAS with protection, i.e., patients no longer receive medical management only; (iii) Costs related to medical management. All patients (including those who undergo CEA or CAS with protection) receive the same medical management. Therefore, MM (CAS with protection, CEA, or medical management alone) does not contribute to incremental differences in cost across treatment alternatives; (iv) Pre-procedural costs. Costs incurred during the time between diagnosis and treatment do not differ across treatment alternatives
These assumptions warranted performance of a cost-minimization analysis. Table 5 presents a breakdown of per patient costs for CEA and CAS with protection. Cost and utilization information were derived from published literature, with one exception (16). Filter device costs were obtained from one of the manufacturers (Cordis). All costs identified are in U.S. dollars. The per case treatment cost for CAS with protection was found to be slighter greater than that for CEA (582 US). This finding was primarily due to higher equipment-related costs.
DISCUSSION
In high-risk patients with severe CAD, the evidence suggests that CAS with cerebral protection may offer a safe and efficacious alternative to CEA, reducing the risk of embolic peri-procedural complications associated with CAS to acceptable levels, i.e., levels equivalent to those for CEA. Clinicians could, therefore, soon be faced with the need to modify their own practice patterns in response to patient awareness of and preference for the minimally invasive, nonsurgical option. Moreover, in patients who are not able to undergo surgery, it may no longer be appropriate to offer medical management only, given the availability of a well-tolerated, more beneficial approach.
The anticipated increased demand for CAS with cerebral protection could significantly impact the health care system. Any long-term savings of preventing a stroke may be overshadowed by the need to revisit budgets and reallocate resources to cover the procedural costs. In addition, the number of specialists currently performing CAS within a particular health region may no longer be enough to handle patient volumes. Nonetheless, without detailed information on the utilization of CEA and CAS with protection in Canada or abroad, such potential resource and organizational implications are merely speculative.
The economic implications of filter devices depend primarily upon the number of patients likely to receive CAS with filter protection Therefore, to ensure that maximal health outcomes are efficiently achieved with filter protection devices, it is recommended that their introduction into routine clinical settings be controlled, i.e., made available with restrictions, e.g., limited to use in select patient populations most likely to benefit from CAS with filter protection (high-risk, high-grade CAD patients), performed at specific sites by experienced specialists adhering to predefined procedural protocol, and so on.
CONTACT INFORMATION
Devidas Menon, PhD (menon@ualberta.ca), Professor, Public Health Sciences, University of Alberta, 13-103 Clinical Sciences Building, Edmonton, Alberta T6G 2G3, Canada
Tania Stafinski, MSc (tstafinski@ihe.ca), Research Associate, Institute of Health Economics, #1200, 10405 Jasper Avenue, Edmonton, Alberta T5J 3N4, Canada
