Assessing the Learning Curve of CAS
Lessons learned from 246 carotid artery stenting procedures.
This study evaluated the learning curve and clinical outcome of carotid artery stenting (CAS) with routine use of a cerebral protection device. The authors report a definite procedure-related learning curve, as evidenced by the reduced number of procedure-related complications, fluoroscopic time, and contrast volume that occurred as a result of an increase in physician experience. The procedural success is also enhanced partly by endovascular device refinement and an improved anticoagulation regimen. Successful outcome of CAS can be achieved once physicians overcome the initial procedure-related learning curve.
Since the first reported cases of carotid artery angioplasty by Kerber et al in 1980,1 the popularity of percutaneous carotid intervention has steadily risen due in part to the perceived benefits of the less-invasive nature of the procedure, less procedural discomfort, and faster convalescence when compared to carotid endarterectomy (CEA). In the past several years, endoluminal devices used in CAS have been refined to improve the treatment success and reduce procedural complications. In particular, various neuroprotection devices have been introduced in clinical practice to minimize the risk of plaque embolization during the CAS procedure.
Several prospective clinical studies that compared the clinical efficacy of CEA and CAS have demonstrated a similar or even superior outcome in percutaneous carotid interventions in high-risk surgical patients.2-6 The outcome of the CAVATAS (Carotid and Vertebral Artery Transluminal Angioplasty Study) trial revealed equivalency between the two treatment modalities with regard to neurologic complications and freedom from stroke at 3 years.7 The SAPPHIRE (the Stenting and Angioplasty with Protection in Patients at High Risk for Endarterectomy) trial, which randomized high-risk surgical patients to either CEA or CAS with neuroprotection, showed a superior efficacy of CAS with significantly lower combined stroke and complication rates at 30 days when compared to the CEA cohorts.5,6 Further analysis of the data showed persistent superior clinical efficacy of CAS compared to CEA in several patient subgroups at 1-year follow-up.5,6
ONE CENTER'S EXPERIENCE
Given the remarkable clinical outcome of CAS in these studies, it is noteworthy that physicians who participated in these trials had to accumulate certain experience in endovascular carotid intervention before becoming eligible as physician investigators. The minimum number of carotid interventions required for physician investigators varies in these trials, and these requirements ranged from 15 to 25 CAS procedures.7,8 The main impetus of such an entrance requirement is to ensure physician investigators have overcome the initial learning curve of CAS and have presumably acquired sufficient experience as they participate in a clinical trial. Despite increasing numbers of studies that reported improved procedural success of CAS performed in large centers by experienced interventionists, literature remains scarce regarding the learning curve of the CAS procedure. Therefore, this study was undertaken to evaluate the effect of the learning curve of CAS with neuroprotection over a 32-month period at a single academic institution. Specifically, we analyzed whether increasing experience of the procedure affected the procedure-related complication rate and treatment outcome.
PATIENTS AND METHODS
Clinical data of 246 consecutive CAS procedures in 222 patients who underwent percutaneous CAS with neuroprotection were retrospectively reviewed. These patients were treated over a 47-month period ending in August 2005. All CAS procedures were performed by vascular surgery staff physicians at Baylor College of Medicine-affiliated hospitals. Treatment indications were based on high-risk criteria adapted from a recent consensus report.9 All patients underwent preoperative carotid duplex ultrasound that identified high-grade internal carotid artery (ICA) stenosis (³70% luminal narrowing), followed by carotid angiography to confirm the carotid lesion. For the purpose of comparison, analysis was performed in five sequential cohorts of groups (1 to 5), with each group containing 50 consecutive CAS patients.
The patient is given clopidogrel (75 mg/day) and aspirin (81 mg/day) beginning 3 days prior to the intervention. After the stenting procedure, clopidogrel was continued for 3 months; aspirin was continued for life. Prior to June 2002, 54 patients in group 1 received an intravenous (IV) heparin bolus (100 U/kg) to achieve systemic anticoagulation during the carotid intervention. After June 2002, an intraoperative anticoagulation regimen of an IV bivalirudin bolus (0.75 mg/kg, Angiomax, The Medicines Company, New York, NY) followed by an infusion rate of 2.5 mg/kg per hour was used in all remaining CAS procedures. At the completion of the carotid stenting, IV bivalirudin was discontinued.
Cerebral Protection Device
A cerebral protection device was used in all patients undergoing carotid stenting procedures in our study. Patients received either one of the following three cerebral protection devices: (1) the PercuSurge Guardwire distal occlusion balloon (n=22, Medtronic AVE, Santa Rosa, CA), (2) the FilterWire EX system (n=102, Boston Scientific Corporation, Natick, MA), or the FilterWire EZ system (n=76, Boston Scientific Corporation), or (3) the Rx Accunet distal protection system (n=46, Guidant Corporation, Indianapolis, IN).
All interventions were performed percutaneously via a femoral artery access under local anesthesia. The technical detail of our CAS procedure was previously reported.10,11 Briefly, a 6-F introducer sheath was first inserted into the femoral artery. After an arch aortogram was obtained, selective carotid artery catheterization was next performed using either a SIM2 or JB2 catheter (Boston Scientific Corporation, Natick, MA) over a .035-inch Bentsen guidewire (Boston Scientific Corporation). A guidewire exchange was performed in which a 260-cm, .035-inch stiff Glidewire (Terumo Medical Corporation, Somerset, NJ) was used to cannulate the external carotid artery. Catheter exchange was then performed in which a 7-F, 90-cm carotid guiding sheath was inserted into the distal common carotid artery. A .014-inch guidewire system with the distal embolization device was then used to cross the internal carotid lesion. After the activation of the embolization device, a monorail angioplasty balloon was used to predilate the carotid lesion, if necessary.
A self-expanding monorail carotid stent was then deployed across the internal carotid stenosis. Balloon angioplasty after stent placement may be performed depending on the appearance of the completion angiogram. Completion angiography includes biplanar carotid and cerebral views to document vascular anatomy and exclude cerebral thromboembolism. The cerebral embolization protection device was then deactivated and removed. After the completion of the CAS procedure, a standardized antiplatelet therapy was followed in all patients that included daily oral aspirin (81 mg) and clopidogrel (75 mg) for 1 month. Follow-up carotid duplex ultrasound was performed at 1 month, 6 months, and yearly thereafter.
Data Collection and Analysis
Technical success and postprocedural complications were the primary outcome for this analysis. Technical success was defined as achieving <30% residual stenosis after successful CAS. A procedure-related complication was defined as an adverse event that was related to the procedure, including neurologic, cardiopulmonary, and hemorrhagic complications. In addition to post-CAS complications, procedure-related factors, such as procedural time, contrast load, length of hospital stay, and reintervention were analyzed among five patient groups for potential learning curve. A Cox regression model, in a stepwise procedure, was used to identify the most predictive variables associated with procedure-related complication. Chi-square analysis and paired Student t-tests were performed where appropriate; statistical significance was assumed at P<.05.
No significant differences were noted among these patient groups with regard to variables, including demographic factors, treatment indications, or relevant medical comorbidities. The mean carotid artery stenosis was 84%±10%. Overall, a total of 246 carotid stenting procedures were performed in 222 patients (198 men; mean age 72.6±8.4 years; range, 52 to 85 years). Among them, 180 procedures (73%) were performed for asymptomatic lesions, whereas 66 cases (27%) were performed because of symptomatic carotid artery disease.
Overall technical success was achieved in 241 CAS procedures (98%). Table 1 summarizes the technical results, procedural variables, and treatment complications among four patient groups. Procedural success in the latter three groups was significantly higher than in group 1 (100% vs 94%; P<.05). Technical failure occurred in four procedures, which included three cases in group 1 and one case in group 2. These technical failures were all due to severe tortuousity of the common carotid artery or innominate artery, which precluded safe advancement of the carotid guiding catheter. Based on the experience of these four cases, we have since regarded severe carotid tortuousity as a relative contraindication for the stenting procedure. As a result, all patients in groups 3 to 5 had complete technical success (100%).
No significant difference in length of hospital stay was noted among the different patient groups. The overall mean hospital length of stay was 1.4±1 day. However, patients in group 1 had a significantly longer procedural time and received a larger contrast load than did the latter three groups (Table 1). An actuarial plot of procedural times of all CAS procedures is illustrated inFigure 1. There was a clear downward trend in the procedural time as the CAS volume increased. No difference in cardiopulmonary complications was noted among different groups, as there were seven periprocedural cardiac complications among the four patient groups (NS). Hemorrhagic complication occurred in three cases in group 1, in one case in group 2, and in no cases in groups 3 to 5. This difference was also statistically significant when comparing group 1 to either group 3, 4, or 5 (P<.05).
The overall 30-day stroke and death rate was 2.5% (n=5), which included three post-CAS strokes and one death in group 1 and one post-CAS stroke in group 2. Minor stroke occurred in two patients, which included one patient with transient dysarthria and another patient with hemiparesis. Urgent duplex ultrasonography and angiographic study of the carotid vessels were performed immediately after the onset of neurologic symptoms, thus ruling out any technical defect related to the carotid stents. Both patients had complete neurologic recovery at the time of discharge. This combined 30-day stroke and death rate occurred at a higher rate in group 1 compared to the remaining four groups (P<.05; Table 1). When these adverse events are combined, the overall complication rate in our series was 6.5% (n=16). The complication rates for groups 1, 2, 3, 4, and 5 were 18%, 8%, 2%, 2%, and 2%, respectively. The complication rate was significantly higher in group 1 when compared to each of the remaining three groups (Table 1). Using a Cox proportional regression risk analysis to assess procedural complications, low procedural volume (P=.03) was identified as a predictive variable. When applying this predictive variable in a Cox regression model, an actuarial plot was created to predict procedural complications based on CAS volume (Figure 2).
Follow-Up After Stent Placement
Carotid duplex scans were performed at 1, 6, and 12 months, and yearly thereafter. The mean follow-up period was 19.6 months. During this period, 14 patients (7%) died, which included two from a contralateral stroke, two from cancer, five due to myocardial complication, one from an automobile accident, and four due to an unknown cause. High-grade in-stent stenosis with an 80% or greater diameter reduction developed in seven patients (3.5%) at 6 and 8 months. All of these patients underwent repeated balloon angioplasty of the recurrent in-stent stenosis, which resulted in a successful outcome in all cases. All of these patients remained free of neurologic symptoms without recurrent restenosis during subsequent follow-up.
In contrast to many endovascular peripheral arterial interventions, percutaneous carotid stenting represents a more challenging procedure because it requires complex catheter-based skills utilizing the .014-inch guidewire system and distal protection device. Moreover, current carotid stent devices predominantly utilize the monorail guidewire system, which requires more technical agility, in contrast to the over-the-wire catheter system that is routinely used in peripheral interventions. This percutaneous intervention often requires balloon angioplasty and stent placement through a long carotid guiding sheath via a groin approach. Poor technical skills can result in devastating treatment complications such as stroke, which can occur in part due to plaque embolization during the balloon angioplasty and stenting of the carotid artery. Because of these various procedural components that require high technical proficiency, many early clinical investigations of CAS, which included physicians with little or no carotid stenting experience, have resulted in an alarmingly poor clinical outcome.12-15
In one of the early clinical trials that analyzed the efficacy of CAS, Alberts et al randomized 219 patients with high-grade symptomatic carotid artery stenosis to either CEA or CAS treatment groups.12 All patients in the CAS group received the Wallstent (Boston Scientific Corporation). The 30-day periprocedural stroke and death rate was 12.1% for CAS and 4.5% for CEA (P=.049). The 1-year ipsilateral stroke rate was 3.6% in the surgical group, in contrast to 12.2% in the stent group (P=.022). Due in part to the high complication rate in the CAS patient cohorts, the trial was stopped prematurely by the sponsor. Subsequent data analysis showed numerous methodological flaws of the study, including heterogeneous antiplatelet regimens and potential patient selection bias. More importantly, the study found that a majority of CAS-related complications was clustered around physician investigators with little or no previous CAS experience.12,13 Naylor et al conducted a smaller randomized clinical trial that compared the clinical efficacy of CEA and CAS.16 The study was similarly prematurely halted due to a significantly higher complication rate in the CAS treatment group. Further study analysis suggested that the high incidence of CAS-related complications may be due to the lack of sufficient experience in percutaneous carotid artery interventions of the physician investigator.
The Physician's Experience Does Matter
Numerous clinical reports, including our study, have highlighted the importance of the operator's experiences as a crucial factor in the clinical success of the carotid stenting procedure.15,17,18
Ahmadi et al examined their experience with unprotected carotid stenting and noted that increased neurologic complications were present in their early experience.17 They reported a combined 30-day stroke and death rate of 15% in the first 80 procedures, whereas the rate was decreased to only 5% in the subsequent 240 interventions.17
New et al observed a high complication rate of 7.3% of any stroke or death in 137 carotid stents in their early experience from 1996 to 1997. As they gained more experience in carotid intervention, they noted that the stroke rate significantly declined to 2.8% in 1999.19
Similar findings were also reported by Diethrich et al who noted a 10.9% combined stroke and death rate for their first 110 carotid cases. This rate was in sharp contrast to a rate of 6.2% in their subsequent 179 patients.18
Roubin et al reported an improvement in the treatment outcome as their clinical experience grew in a series of 528 patients undergoing a CAS procedure.20 In their report, the overall 30-day stroke and death rate was 7.4% during the 5-year study period, with an improvement to 4.3% at the end of the study period. Regarding the minor stroke rate, the investigators noted an incidence of 7.1% during the first year, which improved dramatically to 3.1% by the fifth year of the study period.20
In a large multicenter survey of physicians who performed carotid stenting worldwide that included more than 5,210 procedures in 4,757 patients, a learning curve of 50 cases was observed with regard to the success of the stenting procedure. In centers that performed fewer than 50 carotid stenting procedures, the combined stroke and death rate was 6.8%. In contrast, institutions that performed more than 200 procedures had a significantly lower stroke and death rate of <4%.15
Wholey et al reported an updated review of the Global Carotid Artery Stent Registry in 2003 that included 12,254 CAS procedures performed in 53 institutions.14 Similar to their earlier report, these researchers noted a steep learning curve in hospitals with a low CAS clinical volume. In fact, institutions that had performed fewer than 50 distally protected CAS procedures had a 4.04% combined stroke and death rate. In contrast, institutions with high CAS experience, particularly those with more than 500 cases of experience, reported a stroke and death rate of 1.56%.14
Lessons Learned From Our Experience
In our study, a definite procedure-related learning curve was observed and was evidenced by the reduced procedural time and contrast volume as our CAS volume increased. In addition, procedure-related complications similarly declined in the latter four groups in contrast to the group 1 patient cohort. The sharp contrast in the procedural time between the early and recent patient groups underscored the importance of the operator's experience and may be partly responsible for the reduced complication rates. As illustrated in Figure 2, the procedural time in the first 30 CAS cases routinely exceeded 60 minutes. With increased clinical experience, percutaneous carotid intervention with neuroprotection can be performed routinely in less than 40 minutes in our last 140 cases. Prolonged procedural time can potentially increase the thromboembolic risk in cerebral circulation due in part to the occlusive nature of the carotid guiding sheath and repeated catheter manipulation in the carotid artery.
As a physician acquires experience to overcome a learning curve associated with a technically demanding procedure, new skills are learned that invariably facilitate performance and ultimately improve outcome. We learned that one important technique (buddy wiring) greatly reduces the technical challenge of navigating through an angulated or tortuous internal carotid artery. Because most of the neuroprotection devices are made of relatively soft .014-inch guidewire, tracking an angioplasty balloon or stent over a soft .014-inch guidewire can be difficult in an angulated vessel. Using the "buddy-wire" technique, a stiff .018-inch guidewire is first inserted in a tortuous internal carotid artery, which straightens the vessel. This maneuver can greatly facilitate the subsequent placement of the neuroprotection guidewire and eventual interventions. This technique has facilitated device deployment and reduced procedural time in cases of severe internal carotid artery tortuousity.
As we acquired technical skills in our learning curve, we also gained relevant knowledge in clinical literatures and adapted necessary changes to improve our treatment outcomes, which is evidenced in the alteration of our anticoagulation protocol. IV heparin was the anticoagulation agent of choice in the first 54 patients undergoing CAS in our series. The preferred anticoagulant was switched to IV bivalirudin in all subsequent patients. The modification of the anticoagulation resulted in a significant reduction in hemorrhagic complications, which occurred in 6% (three of 50 patients) in group 1 and 2% (one of 50 patients) in group 2, but none in the latter two patient groups. Bivalirudin is a direct thrombin inhibitor with a half life of 45 minutes. As a recombinant hirudin derivative, this agent is given on a weight-based infusion during the CAS procedure. The impetus of this protocol modification stemmed from the recent publication of the REPLACE-2 trial (the Randomized Evaluation in PCI Linking Angiomax to Reduced Clinical Events).21,22 In this study, more than 6,000 patients undergoing coronary intervention were randomly assigned to receive either IV bivalirudin or heparin plus Gp IIb/IIIa blockage. The study found that clinical outcomes at 30 days between the two groups remained similar, except that significantly lower major adverse events and hemorrhagic complications occurred in the bivalirudin cohort than the heparin cohorts. At 6-month follow-up, there was a trend toward decreased mortality in the bivalirudin group compared to those who received heparin and Gp IIb/IIIa blockade.22
Subspecialty Documents on Clinical Competence of CAS
Because a learning curve is a well-recognized phenomenon in endovascular carotid intervention, many subspecialty societies have proposed various credentialing guidelines for physicians to gain the necessary clinical competence to perform this procedure. A joint statement published by the American Society of Interventional and Therapeutic Neuroradiology, the American Society of Neuroradiology, and the Society of Interventional Radiology, proposed that a physician should perform at least 200 cervicocerebral angiograms and 10 supervised CAS procedures to acquire the necessary technical competence to perform the CAS procedure.23,24 Another document was recently published jointly by several societies with regard to the clinical competence of CAS, which included the Society for Cardiovascular Angiography and Interventions, the Society for Vascular Medicine and Biology, and the Society for Vascular Surgery.25-27 This document recommended that a physician should perform at least 30 cervicocerebral angiograms and 25 CAS procedures, with at least half as primary operator, to obtain the minimum clinical competence to perform carotid interventions. In addition to these recommended guidelines, this document emphasizes various means for physicians to acquire the necessary skills and training in CAS, including industry-sponsored courses and carotid simulation training modules. The beneficial role for physicians to acquire the necessary hands-on skills using a carotid stenting simulation system has been well documented in many reports.28-30 The constant refinements of both software and hardware by the simulator manufacturers have already resulted in many simulation systems that provide an extremely realistic and practical training tool for carotid intervention.
Our study demonstrated that there is a definite learning curve associated with CAS with neuroprotection device placement. With many studies documenting an equivalent clinical outcome of CAS when compared to CEA, it is conceivable that this catheter-based intervention will likely replace the traditional surgical therapy in the near future. Physicians with experienced endovascular skills, such as interventional cardiologists or radiologists, are well suited to perform this procedure once they overcome the initial learning curve. Vascular surgeons who do not have the fundamental catheter-based skills, however, must take an active role to acquire the necessary skills to learn and perform this intervention safely. Interventionists who wish to perform percutaneous carotid stenting must be cognizant of a learning curve in this procedure. As physicians gain increased experience in CAS, they not only can overcome this learning curve, but also undoubtedly will provide safe and effective endoluminal therapy to patients with carotid bifurcation disease.
Peter H. Lin, MD, is Associate Professor of Surgery, Division of Vascular Surgery & Endovascular Therapy, Michael E. DeBakey Department of Surgery at Baylor College of Medicine in Houston, Texas. He has disclosed that he has no financial interest in any product or manufacturer mentioned herein. Dr. Lin may be reached at (713) 794-7895; firstname.lastname@example.org.
Wei Zhou, MD, is Assistant Professor of Surgery, Division of Vascular Surgery & Endovascular Therapy, Michael E. DeBakey Department of Surgery, at Baylor College of Medicine in Houston, Texas. She has disclosed that she has no financial interest in any product or manufacturer mentioned herein. Dr. Zhou may be reached at (713) 794-7892; email@example.com.
Panos Kougias, MD, is Assistant Professor of Surgery, Division of Vascular Surgery & Endovascular Therapy, Michael E. DeBakey Department of Surgery, at Baylor College of Medicine in Houston, Texas. He has disclosed that he has no financial interest in any product or manufacturer mentioned herein. Dr. Kougias may be reached at (713) 794-7892; firstname.lastname@example.org.
Hosam El Sayed, MD, is Assistant Professor of Surgery, Division of Vascular Surgery & Endovascular Therapy, Michael E. DeBakey Department of Surgery, at Baylor College of Medicine in Houston, Texas. He has disclosed that he has no financial interest in any product or manufacturer mentioned herein. Dr. El Sayed may be reached at (713) 794-7892; email@example.com.
Alan B. Lumsden, MD, is Professor of Surgery, Division of Vascular Surgery & Endovascular Therapy, Michael E. DeBakey Department of Surgery at Baylor College of Medicine in Houston, Texas. He has disclosed that he has no financial interest in any product or manufacturer mentioned herein. Dr. Lumsden may be reached at (713) 794-7892; firstname.lastname@example.org.