Training, Competency, and Credentialing Standards for Diagnostic Cervicocerebral Angiography, Carotid Stenting, and Cerebrovascular Intervention

Risks of Diagnostic Cervicocerebral Angiography

Stroke is recognized as the most disabling and costly of all medical conditions (2). Stroke is also the most feared of all iatrogenic medical and procedural complications. The risk of procedure-induced stroke may be a reason not to recommend the test for many physicians, and contributes to the reluctance of some patients to undergo the procedure (3–6). For medical and ethical reasons, any procedure that has “stroke” as a defined risk should be performed only by medical professionals with appropriate training and experience.

The risk of permanent neurological deficit as a result of diagnostic cerebral angiography is considerable and ranges from 0.3–5.7% (5, 7–20). Experienced neurovascular specialists may have complication rates lower than 1% (20). There is additional risk of temporary neurological deficit ranging from 0.3–6.8% with, on average, a 2–3 fold increased risk of temporary as compared to permanent neurological deficit (7–20). Patients with atherosclerotic cerebrovascular disease as manifested by neurological symptoms (ipsilateral transient ischemic attack [TIA] or stroke) have a 2–3 fold higher risk of stroke from diagnostic cerebral angiography (0.5–5.7% risk of permanent deficit) as compared to asymptomatic lesions (0.1–1.2% risk) (5–10, 15–20). In one study, 1000 consecutive patients undergoing diagnostic cerebral angiography were assessed for procedure-related neurological deficits (5). The overall stroke rate was 1%. However, 9 of the 10 patients experiencing neurological complications had a history of prior stroke or transient ischemic attack and the tenth had an “asymptomatic” bruit (5). Therefore the highest level of practitioner training should be required for patients with prior symptoms, who are at highest risk for angiographic complications.

Operator experience as measured by decreased complications and decreased fluoroscopy time necessary for the exam improves in a linear fashion up to 100 cases (10). Analysis of the trainee learning curve suggests that 200 exams are necessary for a physician to become a competent and secure examiner of the carotid and intracranial vasculature (10). Operator risk factors for angiographically produced ischemic complications (temporary and/or permanent stroke) are well known and include increased procedure and fluoroscopy time, increased number of catheters used, and performance of arch aortography (6–8). Performance of arch aortography may lead to greater numbers of emboli thus leading to higher procedure complication rates than selective carotid angiography and is not infrequently performed by less well-trained practitioners (8, 21). All of the above-mentioned factors, including procedural time and multiple catheter use, are not independent and are typically related to inexperience and lack of specialized training in the cervicocerebral circulation (8, 12). The effect of training and experience, and/or lack thereof, was clearly shown in a 5000-angiogram analysis that demonstrated that fellowship-trained specialists have fewer neurological complications (0.5%) than even experienced angiographers (0.6%), and both have far fewer complications than trainees under supervision (2.8%) (7, 18, 19). In the Asymptomatic Carotid Atherosclerosis Study (ACAS), the rate of stroke as a complication of diagnostic cerebral angiography was approximately 1.2% (17). This may be greater than the actual risk of stroke caused by the stenosis itself for many patients with asymptomatic stenosis (17). Indeed, this fact has led some vascular surgeons to suggest that diagnostic cervicocerebral angiography even when performed by well-trained neurovascular specialists may be too dangerous for the indication of asymptomatic carotid artery stenosis (22). However, more recent data has confirmed that the rate of stroke during routine diagnostic cerebral angiography when performed by appropriately trained and experienced neurovascular specialists is less than half the rate reported in ACAS (20).

Clinically obvious stroke may be the tip of the iceberg regarding complications of cervicocerebral angiography. “Silent” neuropathological sequelae of cerebral embolism are even more common than overt, clinically demonstrable neurological complications (20, 21, 23–25). The fact that thromboembolic occurrences may be “silent”, yet still represent serious pathologic brain damage has recently been described in two magnetic resonance imaging (MRI) studies where diffusion weighted pulse sequences ideal for detecting small infarcts were obtained after angiography (23, 24). In one study, small new areas of brain infarction without overt clinical correlates were identified in 25% of 66 patients after diagnostic cerebral angiography (23). Detection of apparent embolic insults by MRI was more common in cases with longer fluoroscopic/procedural times (p <0.01) and was associated with the use of multiple catheters (p=0.02) (23). Both of these parameters have been shown to be associated with sub-optimal training and experience (24). “Subclinical” infarcts have been shown to result in cognitive deficits on neuropsychological testing after endarterectomy as well as carotid artery stenting (25). Similar procedural injury to the heart has been extensively documented secondary to coronary interventions by measurements of elevations in troponin levels (so-called troponin “leak”) and constitutes justification for the current stringent training standards for coronary intervention (26, 27).

In addition to the technical risks of cerebrovascular procedures, there is also a risk of misdiagnosis if images are not interpreted correctly. This fact justifies formal and adequate cognitive training related to neurological and neurovascular anatomy, neurodiagnostic imaging, and neuro-pathophysiology. Physicians must be able to accurately identify stroke and TIA etiologies and evaluate traumatic and/or atherosclerotic lesions and inflammatory conditions of the central nervous system. Evidence from numerous studies of coronary angiography performed by trained cardiologists demonstrates errors between observers’ assessments ranging from 15% to 45% for evaluating essentially only one variable, ischemic vascular disease (28). The ramifications of inter-observer variation are considerable. If readings are erroneous, some patients will undergo interventional procedures unnecessarily, others might be denied an essential treatment, while still other patients may have pathological findings that are totally unrecognized (28). The implications of this degree of variability for patients with cerebrovascular conditions are significant when considering that physicians may be performing and interpreting cervicocerebral angiography outside of their primary specialty training and may then be performing interventions that have stroke as a defined potential risk. Even if a cervicocerebral arteriogram is performed solely for assessment of extracranial carotid occlusive disease, unexpected findings (vasculitis; congenital vascular malformations; tumors; mass effects; embolic complications; acute, subacute, or chronic dissection as opposed to atherosclerotic disease; atherosclerotic disease; aneurysms; pseudoaneurysms; arteriovenous fistulae; etc.) require extensive neurodiagnostic and neuroangiographic knowledge and interpretive skills which can only be obtained with appropriate formal training.

Risks of Cervicocerebral Interventional Procedures

Endovascular interventions carry a higher risk than diagnostic angiography in all vascular beds. The American College of Cardiology (ACC) has recognized this by requiring physicians to complete diagnostic coronary angiography training prior to beginning interventional coronary training (29). The risk of elective carotid stenting is greater than the risk associated with elective coronary intervention, which is typically less than 2% for emergency coronary artery bypass surgery and less than 2% for death (30, 31). Randomized controlled trial data indicate stroke and death rates for carotid stenting ranging from 4.4% to over 12% at 30 days, with a one-year stroke and death rate of up to 12% (32–41). MRI examinations demonstrate detectable ischemic lesions in 22% to 29% of brains after carotid stenting (42, 43). Additionally, a significant learning curve for carotid stenting has been clearly documented (44).

Potential benefit from “embolism protection” devices might render carotid stenting safer than is currently documented, but procedural stroke and death rates still range from at least 2.8% in one registry to over 6% at 30 days in other unpublished registries for both asymptomatic and symptomatic patients (34, 36, 37, 40). Indeed, in two randomized controlled trials comparing stent procedures with “protection” and with “no-protection”, there was conflicting evidence concerning protection, with one trial indicating no difference and the other actually demonstrating worse outcomes “with protection” (45–47). Possible efficacy of “protection” devices has been demonstrated in at least one registry, in the carotid stenting arm of an endarterectomy versus stenting trial, and in a review article (40, 48, 49). Therefore, for carotid stenting, the conflicting proof of efficacy for protection devices, proven failure to eliminate all complications including stroke or death, and demonstrated patient risk greater than elective coronary intervention, for example, reaffirms that carotid stenting be performed only by individuals with sufficient cognitive neuroscience knowledge coupled with sufficient training and experience and subsequent excellent procedural technique, as described herein.

Cervicocerebral intervention not only includes carotid artery and extracranial angioplasty and stenting but also intracranial angioplasty and stenting as well as other therapies. The risks of neurological complications from intracranial angioplasty and stenting and cerebral aneurysm coiling are substantial. The reported neurological complication rate for intracranial angioplasty and stenting ranges from 5% in 30 days to 36% (50–59). A significant learning curve has been demonstrated for coiling of cerebral aneurysms and the reported neurological complication rate ranges from 5% to 14% (60–64). Similar to the findings in carotid stenting, diffusion-weighted MRI reveals a higher rate of distal embolization associated with this procedure (up to 61%) than overt symptoms; many of the emboli are “silent” (21, 23, 24, 65).

Introduction

Official standards of training for all specialties have existed for over a quarter century, are the hallmark of medical licensure, board examinations and residency programs, individual physician privileges and hospital credentialing, and are recognized as vital by the Accreditation Council for Graduate Medical Education (ACGME), the Federation of State Medical Boards of the United States, Inc., the American Board of Medical Specialties (ABMS), and the National Board of Medical Examiners® (NBME®) (66–68). Furthermore, continuing assessment of competence is mandated by the Centers for Medicaid and Medicare Services as well as state medical licensing boards in the form of Continuing Medical Education (CME) credits (69–71). The Joint Commission on Approval for Healthcare Organizations (JCAHO) is working with two other accrediting organizations, the National Committee for Quality Assurance and URAC (formerly known as the Utilization Review Accreditation Commission), on coordinating and aligning patient safety standards (72–74). JCAHO has established guidelines for primary stroke centers based on Brain Attack Coalition recommendations that include quality of service standards for diagnostic cervicocerebral angiography (75). The Brain Attack Coalition has also established guidelines for Comprehensive Stroke Centers that mandate cognitive and technical neurovascular training and expertise in order to perform carotid stenting (Alberts MJ, Latchaw RE, Selman WR, et al. Recommendations for Comprehensive Stroke Centers: A Consensus Statement from the Brain Attack Coalition. Submitted for publication.).

Training guidelines for diagnostic arteriography and endovascular intervention are necessary for optimal and safe patient care and have been formulated and officially stated by numerous medical societies, including the American Heart Association (AHA), the ACC, the Society for Vascular Surgery (SVS), the Society of Interventional Radiology (SIR), the American Society of Neuroradiology (ASNR), and the American Society of Interventional and Therapeutic Neuroradiology (ASITN) (76–98). These AHA, ACC, SVS, SIR, ASNR, and ASITN guidelines mandate at least 100 diagnostic angiograms regardless of the vascular bed. The fact that there are varying degrees of difficulty for certain procedures and that these procedures thus impart associated degrees of risk to the patient has also been specifically recognized and summarized by the ACC (79). For example, in recognition of the critical nature of certain catheter based procedures, the ACC has published the Revised Recommendations for Training in Adult Cardiovascular Medicine Core Cardiology Training II statement (COCATS 2) (29). In addition to the required minimum 24 clinical months of training by COCATS 2, diagnostic coronary catheterization mandates a minimum of 8 dedicated months in a cardiac catheterization laboratory during training in the pathophysiology and treatment of heart disease with specific requirements for approved supervised training on at least 300 diagnostic coronary angiograms before a practitioner is judged competent for credentialing purposes (29). This same concept is at least as important when dealing with the cerebral vasculature and the performance of cervicocerebral angiography.

The ACC has determined that cognitive training about the pathophysiology of the heart in addition to credentialing in diagnostic coronary angiography is a prerequisite for training in coronary intervention (80, 84, 86, 87). Furthermore, in addition to the core 24 month training period and 300 diagnostic coronary angiograms, the ACC recommends a full 20 months of supervised cardiac catheterization lab training with at least 250 supervised coronary stent procedures as the minimum acceptable requirements before a practitioner is judged competent to perform coronary interventions (88–92). The ABMS has not only affirmed that high degrees of training are necessary for appropriate and safe cardiac patient care but acknowledged this high level of achievement in the form of a Certificate of Added Qualification (CAQ) for Interventional Cardiology (99). These same principles are necessarily as crucial for the performance of interventional procedures relating to the cervicocerebral vasculature, including carotid stenting.

Existing Standards

Knowledge Necessary For Cerebrovascular Intervention

Our collaborative neuroscience societies, in agreement with the principles espoused in the ACC COCATS 2, recognize the necessity of three components of adequate training for competency to perform cervicocerebral diagnostic and interventional procedures: 1) formal training which imparts an adequate depth of cognitive knowledge of the brain and its associated pathophysiological vascular processes, including management of complications of endovascular procedures, 2) adequate procedural skill achieved by repetitive supervised training in an approved clinical setting by a qualified instructor, and 3) diagnostic and therapeutic acumen, including the ability to recognize and manage procedural complications, achieved by studying, performing and correctly interpreting a large number of diagnostic procedures with proper tutelage. Just as with diagnostic coronary angiography and coronary intervention, extensive knowledge of the brain and the ability to correctly interpret a cervicocerebral angiogram is the prerequisite and foundation for the technical performance of cervicocerebral angiography. The ability to adequately assess the array of diagnostic imaging studies of the brain with adequate knowledge of the numerous pathophysiological possibilities is a necessary attribute of any practitioner who would perform cervicocerebral procedures, irrespective of the primary specialty of the practitioner.

Although interpretative skills of imaging are essential, clinical cognitive skills related to the epidemiology, diagnosis, and management of patients with cervicocerebral vascular disorders are the sine qua non of quality patient care, safety, and treatment selection. All major industry and National Institutes of Health (NIH) sponsored trials related to carotid stenting and cervicocerebral interventions, including asymptomatic, symptomatic and high surgical-risk patients, have required an independent assessment by a board-certified neurologist. This assessment includes documented competency in performing a complete neurological evaluation including the NIH Stroke Scale. Consequently, we not only endorse this principle in general practice, but also mandate adequate training for all neuroendovascular practitioners that encompasses knowledge of stroke syndromes and includes formal training and competency in the NIH Stroke Scale.

Competence in recognizing any procedural complication and being able to offer the most appropriate treatment is one of the basic goals of adequate formal training, particularly concerning cervicocerebral angiography and/or intervention. This would include the ability to recognize clinical intra- or post-procedural neurological symptoms as well as pertinent angiographic findings and the proper cognitive and technical skills to offer the most appropriate therapy. While this therapy might entail intracranial endovascular rescue, it might also entail optimal hemodynamic management necessitating sufficient clinical neurointensive skills.

Our collaborative neuroscience societies recognize that practitioners from a variety of backgrounds may currently have or wish to develop endovascular skills. Our consensus is that a minimum amount of formal cognitive training specifically related to stroke and cerebrovascular disease is essential for any physician to perform diagnostic cervicocerebral angiography and interventional procedures. Therefore, in addition to procedural technical experience requirements, a minimum of 6 months of formal cognitive neuroscience training in an ACGME-approved training program in radiology, neuroradiology, neurosurgery, neurology, and/or vascular neurology is required. This minimum formal training applies to all practitioners who wish to be credentialed to perform diagnostic cervicocerebral angiography and/or cervical carotid interventions, including practitioners from specialties with or without dedicated training in clinical neuroscience as part of their ACGME-approved residency programs.

Augmentation of Training

Simulator training has been shown to be of benefit in limited medical applications (105–112). At the present time, appropriate formal training and experience in clinical cervicocerebral angiography and intervention in an approved clinical training program has no adequate substitute in contemporary medical practice, but future trainees may benefit from added training on medical simulators. At the present time, simulator equipment is neither perfected nor validated for training purposes concerning the cervicocerebral vasculature, but it is anticipated that eventually these technologies may offer up to, but not greater than, 20% of the required training experience in procedural technique. Our collaborative societies, consistent with ACGME training standards and the ACC training standards (COCATS 2), emphasize that industry-sponsored seminars, continuing medical education (CME) coursework, and self-taught learning are insufficient for credentialing related to diagnostic cervicocerebral angiography, extracranial interventions, intracranial interventions, or carotid stenting.

Maintenance and Assurance of Continuing Quality of Care

Procedures that have stroke as a defined potential risk require the highest level of competency. Proficiency is maintained by lifelong continuing medical education as well as continuing performance of cases with adequate success and outcomes with minimal complications. Quality Assurance and continuing improvement are necessary for high quality healthcare regardless of which discipline might be involved in treating patients. The quality improvement process is a patient oriented process, designed to ensure a baseline level of quality and predictable outcomes, and represents in many ways a safety net for the credentialing process. A post-hoc quality assurance process is no substitute for adequate and appropriate physician training leading to acceptably skilled practitioners suitable for credentialing. A quality assurance process should confirm that procedures are performed for appropriate indications with rates of success and complications that meet acceptable standards. Such Quality Improvement standards have been published for diagnostic cerebral angiography as well as extracranial carotid stenting (77, 82, 95, 113). Such standards are necessary for quality assurance for procedures of such considerable consequence. The outcomes required by these standards should be achieved both during the training cases and following granting of credentials in order to ensure maintenance of competence. At this time there is insufficient information to know if maintenance of competency requires annual performance of specific numbers of cases, but data from other vascular interventional procedures such as coronary stenting, coronary artery bypass grafting, and carotid endarterectomy indicate that, in general, greater experience confers better outcomes (114–116).