Rethinking Training and Distribution of Vascular Neurology Interventionists in the Era of Thrombectomy
See related article, p 2042
Vascular neurologists (VNs) now have 2 powerful tools to improve outcomes after stroke, intravenous tissue-type plasminogen activator (r-tPA) and endovascular thrombectomy (ET). Among the many common aspects of both treatments, the fastest possible intervention after stroke onset emerges to be paramount.1–3 With r-tPA, it has taken 20 years from approval to be accepted, the subspecialty of VN to be born and trained, and systems of care implemented to speed treatment by improving Emergency Department and pre-hospital management. In contrast, ET based on the 5 positive landmark randomized trials reported in the past 2 years4–8 has been accepted much more abruptly. Yet, the same sort of changes in streamlining systems of care and training and distribution of VN expertise are as necessary for ET as they were for r-tPA to deliver ET as rapidly as possible. The abruptness of this ET revolution, and the redirection of care and resources required, have found the VN community unprepared. Dramatic changes in VN training and distribution are necessary to accommodate the fastest possible ET intervention. Our 3-fold premise to be explored below is that (1) treatment of ET candidates is multifaceted and optimally should include the expertise of a VN, (2) the VN community needs to rethink and retool its workforce to accommodate the most rapid and widespread ET treatment possible, and (3) this should by necessity include training and more widely distributing substantially more VNs to carry out ET.
The authors acknowledge that what we propose as a Comment and Opinion diverges from the current standard management, which is to concentrate ET expertise into a few hands at Comprehensive Stroke Centers (CSCs) and have the patient brought to Mecca. Initially, the same model was proposed for r-tPA, but the top 3 enrolling sites in the National Institute of Neurological Disorders and Stroke (NINDS) r-tPA for Acute Stroke trial used the commando approach where the stroke teams were physically deployed from the stroke team hub to enroll and treat patients at community hospitals. Using this approach, 50% were treated within the first 90 minutes from the time last seen normal. Gradually it became clear that faster patient care to provide r-tPA administration could be best served not only at CSCs but also closer to the patient at primary and stroke-ready hospitals. This multilevel stroke center concept has been accompanied by efforts to shortcut unnecessary ED delays, distribute expertise more widely by training more VN and leveraging remote expertise via telemedicine, and more recently speed treatment even more dramatically by using Mobile Stroke Units. The impact of these measures has been and continues to be evaluated through registries and comparative effectiveness studies.9,10 For ET, we propose a similar model, except to do it more quickly rather than waiting 20 years.
We suggest that optimal employment of ET in the US requires training and more widely distributing a contingent of a new VN interventionist able to carry out r-tPA, ET, and other aspects of pre- and post-acute management. This is not meant to exclude training and participation of our neurosurgical and neuroradiology colleagues in ET. We support the collaborative efforts of all these specialties to develop common training, guidelines, and research efforts. Furthermore, a wider distribution of ET expertise by neurosurgeons and neuroradiologists outside of the CSC Mecca would be welcome and hopefully will be encouraged by the leadership of those specialties. As senior VNs, however, we hope to awaken our own specialty to what we feel is needed to provide the best care for our patients with acute stroke. For the reasons we will outline, we believe that optimal ET management in the future will benefit from more VN training in ET.
Once the measures we outline below are implemented, registries and comparative effectiveness studies should compare outcomes—including number of patients treated, percent recovering, costs, and patient satisfaction—to the current model of centralized care.
Optimal Management of Patients With Stroke Involves More Than ET and Benefits From the Expertise of VNs
The results of 9 randomized trials including almost 7000 patients1 confirm the initial findings of the NINDS study despite lingering doubts among a small cadre of vocal non-VN physicians.11 r-tPA is effective across all stroke subtypes and severity, regardless of patient age and other demographics. Treatment response is determined mainly by how fast the drug is given after symptom onset. The vast majority of patients with AIS present with National Institutes of Health Stroke Scale scores <12 and do not harbor clots amenable to ET.12 Even if future studies show that r-tPA can be omitted in some patients going for ET, r-tPA will remain a critical aspect of acute stroke treatment.
The benefits of r-tPA clearly outweigh the risks of bleeding for appropriately selected patients. Because the risks of bleeding are real, and the cost of the drug is high, true strokes must be distinguished from stroke mimics. Hence, optimal patient care requires the skill of physicians trained in stroke recognition and r-tPA administration. This certainly is a skill that other specialists in Emergency Medicine, Internal Medicine and Neurosurgery, and other healthcare professionals such as nurse practitioners can learn and do well. Yet, we propose that, optimally, VNs carry out or at least participate in r-tPA treatment decisions. For these reasons, training in the use of r-tPA is an essential and required aspect of all VN fellowship programs and for all VN fellows.
Similar to r-tPA, successful response to ET is strongly influenced by how fast the artery is recanalized after symptom onset. In addition, appropriate selection of ET candidates is critical and can be facilitated by the clinical expertise of a VN. For instance, a patient with subtle brain stem signs may have a low National Institutes of Health Stroke Scale score and harbor a basilar artery occlusion, and a patient who harbors a middle cerebral artery occlusion may have fluctuating cortical signs/symptoms as the only clue. On the other side of the coin, although still a matter of investigation, patients with large-vessel occlusion may have an unfavorable imaging profile or such significant comorbidities as to make them poor candidates for intervention. Finally, preliminary results from the DAWN trial (Clinical Mismatch in the Triage of Wake Up and Late Presenting Strokes Undergoing Neurointervention With Trevo) suggest that carefully selected patients who present late may respond favorably to ET if they have a favorable imaging profile.
The benefits of both r-tPA and ET can be eroded if post-acute care is suboptimal. AIS treatment requires management in specialized Stroke Units or at a minimum, ICU beds for stroke, manned by physician and ICU personnel with appropriate training. This includes training to prevent and recognize acute stroke complications, carrying out accurate and cost-effective diagnosis of stroke cause, designing and implementing appropriate secondary stroke prevention strategies, and beginning mobilization and the road to rehabilitation and recovery.13,14 Decisions about feeding, activity, and antithrombotic therapy must be made—and made correctly—within the first 48 hours after stroke onset. The appropriate diagnostic evaluation must be determined from among a myriad of tests, in particular imaging, that are now available. Results must be carefully evaluated and put into context, while avoiding unnecessary cost and patient inconvenience. The benefit of rapidly applied secondary stroke prevention measures targeting the particular stroke subtype and specific underlying pathology is of unequivocal importance. Timing and choice of rehabilitation strategies must be determined. Mortality rates from stroke continue to decline as a result of these measures. In addition, incorporating these practices to achieve excellent outcomes will become essential to assure reimbursement in the coming era of bundled or capitated payments. Consequently, these aspects of stroke care are also embedded in all VN training programs and an important focus of the education of all VN fellows.
Carrying Out ET Without VNs and Attention to These Other Aspects of Stroke Care Is Not in the Best Interest of Our Patients
Our colleagues in Emergency Medicine, Internal Medicine, Cardiology, Neurosurgery, Neuroradiology, Critical Care, and Physical Medicine and Rehabilitation are important partners. With regards to ET, neurosurgeons and interventional neuroradiologists are skilled in delivering ET and currently partner with VNs in treating such patients and should continue to do so. Although an area of debate, cardiologists might also desire to carry out urgent intracranial ET although only after appropriate and extensive additional training in the unique nuances of cerebrovascular anatomy, physiology, and clinicopathological correlation. However, for the reasons already stated, stroke care is multidimensional. Optimal and even minimally acceptable stroke care in 2017 requires more than ET. VNs are trained and motivated to care for the entire spectrum of stroke management. Ideally, there would be enough VN manpower to oversee the management of every patient with stroke in the United States This is becoming more possible with increased utilization of telemedicine between hub stroke centers and peripheral spokes. Although perhaps care by VNs is not possible for every patient with stroke today, severely affected patients with AIS with large artery occlusions certainly can benefit from the expertise of VNs. In our view, ET should be performed only in settings where such care by VNs is available.
More VNs Should Be Trained to Carry Out ET, and VNs Trained in ET Also Need to Manage the Other Aspects of Stroke Care
The manpower needs to provide ET for all eligible patients with AIS in the United States remain an area of debate and uncertainty. An argument of non-VN specialists such as cardiologists who wish to carry out ET is that there are not enough VNs to carry out ET in community hospitals where they practice and where patients with acute stroke initially present. At the same time, the VNs, neurosurgeons, and interventional neuroradiologists already trained to do ET say that ET should be concentrated in CSCs where there is expertise and large enough volumes to ensure maintenance of skills. We suggest that these arguments for more or fewer practitioners in this new era of ET should be reframed with a fresh perspective on how expertise and practice should evolve to best serve our patients with stroke.
The first argument—that nonstroke-trained specialists should carry out ET outside of CSCs fails for the reasons already articulated. Patients with AIS, in particular those with large artery occlusions, are best managed in an environment where there are VNs. Therefore, an ideal solution to the shortage argument would be to distribute ET-trained VNs more widely and organize our triage systems to rapidly transport patients with AIS to those centers where VNs are located. It is doubtful that there are many community hospitals where interventional cardiologists practice that do not also have access to a VN. Why should not those VNs rather than cardiologists carry out ET along with the other aspects of good stroke care that are part of VN expertise?
ET Capability Must Expand Beyond Our CSC Training Centers
The alternative argument that all ET should be concentrated in CSCs where clinicians trained to do ET can do those interventions with higher volumes-per-interventionist is not ideal because time is brain. Achieving reperfusion quickly is just as important for the success of ET as it is for r-tPA.2,3 At least in the United States, this fact creates the demand for more clinicians trained in ET, and distributing them more widely. Although some countries have small distances and a centralized healthcare system that can mandate patient triage and staffing, this is certainly not the case in the United States. In the entire United States, there are currently 117 Joint Commission on Accreditation of Healthcare Organizations-certified CSCs credentialed to carry out ET (https://www.qualitycheck.org/search/?keyword=comprehensive%20stroke%20center)—slightly over 2 per state. Even if we add CSCs certified by other organizations, many ET candidates live in areas where ET-capable centers do not exist because they are either geographically remote or underserved and, therefore, cannot get transported without substantial delay. We have been unable to identify a publication showing the distribution of VNs in the United States. However, an informal survey of New York and Texas medium-sized cities identified 20 cities in those 2 states alone with >75 000 population, practicing VNs but no CSC or ET capability. This does not include suburban and exurban areas of large cities where VNs practice and where transport to a CSC even by helicopter is associated with substantial delay.
Certainly, we need to support our CSCs, particularly those large training centers where triage systems can concentrate appropriate patients with AIS. It is vitally important that we create such triage systems and support such centers because they are critical to enable the necessary training of new ET VNs. But even in those centers, most neurosurgeons carrying out ET usually do other neurosurgical procedures, and interventional neuroradiologists do other diagnostic procedures. One could argue that the US healthcare system should be better organized to facilitate triage to CSCs. This may be appropriate for urban cores and extremely rural areas. However, a substantial part of the population lives in suburban or exurban areas or medium-sized cities where VNs are already in practice. Most hospitals in these areas are without sufficient neurosurgical volume to support full-time neurosurgical, neurocritical care, aneurysm, and other CSC metrics, but many of them do support viable VN practices.
Our premise is that it is better for a patient in those environments to be treated in the regional hospital stroke centers covered by a VN who does ET than to take the extra ≥2 hours required for transfer to a CSC and then perhaps further delay if the busy interventional VN, neurosurgeon, or neuroradiologist is already occupied on another case. Analagous to the evolution of our approach to r-tPA therapy, and for the reasons we have articulated, we think that regional distribution of ET-trained VNs can be the model for ET as well. We are not proposing that ET be done at centers that cannot provide adequate pre-, intra-, and periprocedural care, but such care does not require a CSC; we propose that ET can be provided at regional stroke centers by VNs who are appropriately trained and who have created a team of nurses, anesthesiologists, technologists, radiologists, and stroke unit personnel to take care of the patient.
Therefore, to provide the most rapid ET treatment to the most patients, we advocate and envision an acute intervention team available 24/7/365 in most stroke centers, primary or comprehensive. In such a scenario, it is unlikely that the interventionist will be occupied 100% of their time by interventions. By staffing the intervention call panel with VNs trained in ET, the medical center gains additional stroke expertise. That benefits the acute patient immediately, during the post-thrombectomy phase and into rehabilitation and ultimately secondary stroke prevention.
We recognize that the model of care we propose is not how the clinical trials demonstrating the efficacy of ET were performed. However, clinical trials are not how clinical care is delivered in real life—that is why post-marketing surveillance is needed for new drugs and for new procedures such as ET. The commando model used in the NINDS trial of r-tPA for Acute Ischemic stroke evolved over time to the distributed, regional model we know today. As we have already stated, once the measures we suggest are implemented, registries and comparative effectiveness studies should compare outcomes to the current model of centralized care. These outcomes should include the number of patients treated, percent recovering, costs, and patient satisfaction.
What Are the Impediments to Training More VNs to Do ET Outside of CSCs?
Experience and practice result in better outcomes, and this is likely the case with ET. Yet, we predict that VNs practicing in regional centers and carrying out ET should be able to achieve the same improved outcomes reported from the positive randomized trials. For example, in the 5 pivotal randomized ET trials,4–8 the mean numbers of cases per 12 months per center were 2.4, 3, 8.2, 9.4, and 22.8. We recognize that the total number of cases done at a given center was probably significantly higher because many patients were likely treated outside of the trials. However, even if we take the 22.8 patients treated with ET in REVASCAT (Randomized Trial of Revascularization With Solitaire FR Device Versus Best Medical Therapy in the Treatment of Acute Stroke due to Anterior Circulation Large Vessel Occlusion Presenting Within Eight Hours of Symptom Onset; the only study where the government mandated a registry of all eligible patients) and add an equal number who were treated outside the trial, the number of patients treated per interventionist would be no greater than what we calculate below considering that the total number treated at each center was usually divided among >1 interventionist.
The authors do not purport to know how many cases and their complexity that need to be performed each year to maintain ET skills in practice. This number should be determined by those experts who carry out the procedures and train the practitioners. However, this number, and how much results differ between centers based on the number of cases performed, should be confirmed by registry studies and real-world experience now that the frequency of ET is spreading, and not set at an unreasonably high level. It is estimated that there are at least 600 000 patients with a new AIS in the United States per year. Based on 2016 data from Memorial Hermann Hospital in Houston, ≈16% of AIS presenting to the emergency department via Emergency Medical Services are ET candidates. Recognizing the selection bias of extrapolating from a major stroke center that benefits from preferential triage of severe stroke patients by Emergency Medical Services, let us assume that nationwide 10%, or 60 000 per year, would be ET candidates. This number is likely to increase based on forthcoming data from the DAWN trial that patients who wake up with stroke symptoms or have uncertain onset time also benefit from ET if they have a favorable imaging profile. Based on a previous review of VN manpower in 2012, by 2016, there should be ≈1500 VNs practicing in the United States15 and ≈80 newly trained VN each year. The same review estimated that by 2016, there would be ≈1200 interventionists already practicing in the United States and 100 newly trained interventionists each year, some of whom would be VN. Let us assume that we start training 50% of VN to do ET. So that means 1200+100+40=1340 interventionists in 2017. That computes to 60 000/1340=45 ET patients for every interventionist. One can quibble with the numbers, but even if we continue to train 50% of our VNs to carry out ET (a highly optimistic and likely unnecessarily high projection), as long as they are distributed more widely so that all eligible patients can access them, 45 cases per year might be a reasonable estimate. Based on the results of the 5 randomized trials, this should be enough to result in good outcomes. It is certainly more than the number of aneurysms clipped or coiled per year by most vascular neurosurgeons and more than the number of aortic valves implanted via the transfemoral route by most interventional cardiologists. At Memorial Hermann Hospital in 2016, 440 patients were treated with r-tPA and 102 received ET evenly divided among 3 interventionists. As we train more VN to carry out ET, and imaging capability spreads in the post-DAWN era, we predict that our ability to recognize and triage ET candidates will increase in parallel, so the number of cases per VN should stay fairly constant for the near future.
What about practical issues such as on-call? A single interventional VN could not be on 24/7 to handle all the ET cases they would do per year. This is the same dilemma faced by any VN in a regional stroke center who has to cover r-tPA call. The solution in many places is to form a 2- or 3-person group to share call. If all these VN were also trained in ET, the schedule would be the same as r-tPA call.
How much training is needed? Appropriate training guidelines have been established by the relevant specialties and must not be shortcut.16 However, although thorough and expert training in all aspect of neurovascular intervention is required, we suggest that it might not be necessary that all VNs carrying out ET outside of CSCs maintain skills to treat aneurysms, AVMs, embolectomies, and other non-ET cases. Those conditions are less time-sensitive, less common, and can be triaged to the specialized CSCs. More of the required Accreditation Council for Graduate Medical Education training year should include study of neurovascular anatomy, emphasis on nuances of existing data and appropriate selection of patients for ET, and preliminary exposure to endovascular techniques. With such added exposure during the Accreditation Council for Graduate Medical Education-approved VN year, sufficient endovascular training could be accomplished within 1 to 2 additional years as recommended in the training guidelines. Finally, smooth and more certain transition from the VN year to the interventional training portion should be coordinated between those in charge of the various phases of the dual training program.
VNs certainly have the skills to carry out ET. We see their outstanding results in our practices every day. Interventionists of any specialty are successful in part because of their innate skill. Like everything else, outstanding results are also because of practice, desire, and creation of a successful team. The expertise of our existing VN colleagues trained in ET attests to the fact this is a skill that can be learned by many VNs with the desire and training to do so. Furthermore, building more ET training into our VN training programs will attract more candidates who have the natural skills and desire to become outstanding interventionists.
So finally, this brings up motivation. In his recently published autobiography, Springsteen17 attributed his success to “DNA, natural ability, study of craft, development of and devotion to an esthetic philosophy, naked desire, and …a furious fire in the hole that just don’t quit burning.” Those same attributes are what most successful clinicians would report. There is no question in our minds that after >3 decades of training VNs both the motivation and skill are to be found in most, if not all, neurologists who select a career in stroke. We just need to create the pathway.
In summary, we would not consider VN training complete without learning how to administer r-tPA and other evidence-based aspects of acute stroke care. Likewise, the authors advocate a larger part in our VN curriculum for vascular anatomy and angiography and that we should create much more ambitious targets for the proportion of VNs trained to carry out ET.
Drs Lyden and Brott reports research funding from National Institutes of Health unrelated to this topic. Dr Grotta reports research funding from the Patient Centered Outcomes Research Institute unrelated to this topic.
The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.
- Received December 21, 2016.
- Revision received March 21, 2017.
- Accepted April 18, 2017.
- © 2017 American Heart Association, Inc.
- Emberson J,
- Lees KR,
- Lyden P,
- Blackwell L,
- Albers G,
- Bluhmki E,
- et al
- Goyal M,
- Jadhav AP,
- Bonafe A,
- Diener H,
- Mendes Pereira V,
- Levy E,
- et al
- Schwamm LH,
- Fonarow GC,
- Reeves MJ,
- Pan W,
- Frankel MR,
- Smith EE,
- et al
- Bowry R,
- Parker S,
- Rajan SS,
- Yamal JM,
- Wu TC,
- Richardson L,
- et al
- Reeves M,
- Khoury J,
- Alwell K,
- Moomaw C,
- Flaherty M,
- Woo D,
- et al
- Dhamoon MS,
- Moon YP,
- Paik MC,
- Boden-Albala B,
- Rundek T,
- Sacco RL,
- et al
- Smith EE,
- Hassan KA,
- Fang J,
- Selchen D,
- Kapral MK,
- Saposnik G
- Leira EC,
- Kaskie B,
- Froehler MT,
- Adams HP Jr
- Lavine SD,
- Cockroft K,
- Hoh B,
- Bambakidis N,
- Khalessi AA,
- Woo H,
- et al
- Springsteen B
- Optimal Management of Patients With Stroke Involves More Than ET and Benefits From the Expertise of VNs
- Carrying Out ET Without VNs and Attention to These Other Aspects of Stroke Care Is Not in the Best Interest of Our Patients
- More VNs Should Be Trained to Carry Out ET, and VNs Trained in ET Also Need to Manage the Other Aspects of Stroke Care
- ET Capability Must Expand Beyond Our CSC Training Centers
- What Are the Impediments to Training More VNs to Do ET Outside of CSCs?
- Info & Metrics
- ET Capability Must Expand Beyond Our CSC Training Centers
- What Are the Impediments to Training More VNs to Do ET Outside of CSCs?
- Info & Metrics