Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
In an iconic scene in ‘The Matrix’, Morpheus gives Neo the choice of taking either the blue pill or the red pill. The blue pill to keep him in his current blissful, fantastical world of illusion and the red pill to shatter that fantasy jolting him into the ‘real world’. The current hype about the potentially explosive growth of endovascular stroke therapy is a blue pill. Neo of course takes the red pill and wakes up to the realities of his time, but will we?
Don't get me wrong, I firmly believe that the recent trials will do for acute ischemic stroke what the International Subarachnoid Aneurysm Trial did for aneurysms. Over a few months we reached milestones that were thought to be many years away, if not unreachable. Suddenly, what we did as neurointerventionalists made sense and was supported by data—not data open to interpretation, but unstinted evidence that appropriately triaged patients will benefit from endovascular revascularization. These trials were the result of a focused, shared objective and collaboration between physicians, industry, and government agencies to assess the cohort of patients best suited for endovascular therapy— that is, those with large vessel strokes.
Working at a center that participated in these and previous studies, I am well aware of the resource-intensive and logistically challenging nature of interventional ischemic stroke trials. About 3 years ago, after much soul searching and review of the evidence, our institution decided to offer endovascular therapy only to patients in a clinical trial. We did this despite our personal convictions and criticisms of the negative endovascular studies because we might be wrong—and if wrong, then doing everything right to discover the truth. Thus all eligible, consecutive patients were enrolled without bias in a trial and we experienced the agony of randomizing relatively young patients with middle cerebral artery occlusions to medical management. That is why the feeling that I had at the positive conclusion of these trials was not dissimilar to that of someone who has been underwater for a while and could now suddenly breathe—big gulps of air. No longer did we have to randomize someone to a treatment that we had previously only thought, but was now proved, to be inferior. We now had ethical and scientific reasons to offer a procedure that we believed would help a patient with stroke.
Why then all the negativity you say? Well, we have achieved what we wished for—but are we ready for it? As the euphoria and excitement settle down are we going to deal with this new found fortune (translate responsibility) with maturity and leadership or are we going to run around like kids in a candy store? According to some, it is a huge candy store and as opposed to falling over each other there will be too much left to rot unless we bring in a lot of people to move it off the shelves. Candy of course is not the right word. I merely use it to denote one extreme if not ugly aspect of an out of control healthcare system that is driven by billboards touting the latest cure to ‘capture the market’, where potential overprojections by analysts and oversupply of physicians run the risk of creating an unrealistic demand. The reason for this cynical preamble is to induce a dose of reality and to highlight some of the challenges in disseminating endovascular stroke care. Our success will be based on an honest and credible appraisal of the current field before projecting its future—a future that is predicated on the real world and an appraisal that requires us to swallow the red pill.
Predicting LVO numbers
Over the past 20 years almost every ischemic stroke presentation or paper that I heard or read began with or contained the statements, ‘almost 800 000 strokes a year’, ‘stroke is the leading cause of morbidity’, and ‘the third leading cause of death’. These figures were often repeated to highlight the seriousness of this disease and to motivate practitioners and hospitals in providing timely care. Thus was launched the era of primary stroke centers, of ‘time is brain’ and of a public awareness campaign to help people recognize the symptoms of stroke and rush to the nearest hospital. We have now reached a stage where these statistics need to be re-examined, if not refined, in the context of large vessel strokes. An easy number exercise that many have performed starting with 795 000 acute strokes, 87% ischemic, 40–50% LVOs, and 25–50% intervention eligibility is to come up with a range of approximately 69 000–170 000 large vessel acute ischemic strokes amenable to thrombectomy. The annual age-adjusted incidence of first-ever ischemic stroke is 179/100 000 for whites and 294/100 000 for blacks.1 Since 77% of the population is white, this yields about 200 new ischemic strokes per 100 000 people. Again, assuming that 40–50% are LVOs gives us 80–100 new large vessel strokes per 100 000 person-years. If we then apply patient eligibility of 25–50% based on imaging and other comorbidities we end up with a range of 20–50 endovascular interventions per 100 000 LVOs eligible for thrombectomy.
Let us compare these numbers with acute coronary interventions, almost all of which are done for ST-segment elevation myocardial infarctions (STEMIs). The estimated annual incidence of myocardial infarctions is 735 000, of which 525 000 are new and 210 000 recurrent attacks.1 Another estimate based on the number of patients discharged with acute coronary syndrome (ACS) from hospitals in 2010 is 625 000, of which 595 000 are for myocardial infarction and 30 000 for unstable angina.1 The percentage of patients with ACS and STEMI is variably reported in different registries and ranges from 29% to 47%.1 Two other registries have reported an annual incidence of 73 and 77 STEMI cases per 100 000, respectively.2 ,3 Regardless of the slight variation, the incidence of STEMIs has been steadily decreasing over the past 20 years to as low as 23%,1 ,3 mainly owing to significant improvement in primary prevention efforts.1 ,3 An average incidence of STEMI of 30% yields about 60 cases per 100 000 person-years (range 40–70).
These are general estimates based on assumptions and your guess may be as good as mine, but these statistics are still useful in putting a cap on the market. These wide ‘ranges’ also highlight the need for accurately coding large vessel strokes and developing central reporting registries. Precise numbers for LVO admissions are thus the foundation upon which to base our development plans.
Projecting thrombectomy growth
A range of 65 000–165 000 new large vessel strokes a year is a debatable if not inaccurate target. Assuming last year there were 10 000 thrombectomies in the USA (the range is 8000–12 000 depending on the information source), it will require a growth rate of 40% a year to reach a target of 30 000 and 80% a year to reach 50 000 in 5 years, respectively. Reaching 65 000 stroke interventions requires a staggering growth rate of 110% a year sustained every year for the next 5 years. That is obviously a bit out of reach but even getting to 65 000 in 10 years requires an annual growth rate of 55%. No procedure in healthcare has shown that kind of growth, and for that matter no industry has reported that kind of annualized return.
Data from the Global Registry Of Acute Coronary Events (GRACE) showed that from 1999 to 2006, primary percutaneous coronary interventions (PCIs) for patients admitted with STEMI grew at about 30% for the first 2 years, 19% for the next 2, and dropped down to an average of 9% for the last 2 years.2 Since then the rate has been declining, especially for non-emergent interventions. One study comparing the trends of primary PCI between 2001 and 2004 reported an overall increase of 24%4; remember this is not annual growth rate but the total increase over 3 years. In another community-based study there was a 69% increase over 15 years.5 It would be wise to keep in mind that even the steepest upslope of the acute coronary intervention era did not top an annual 30% growth rate and that only for the initial years. This is when the American Heart Association (AHA) juggernaut was doing everything to develop infrastructure, raise awareness, and certify centers. What special powers does our field possess, whether physicians or industry to project 40–50% growth rates for LVO interventions even for the first stage?
To define a reasonable target for the next few years we can look at the centers that have been doing stroke interventions for some time, have processes and protocols in place, and are less likely to see major changes in their approach to large vessel strokes. These centers also have relatively less potential for growth as they are already capturing a large percentage of their catchment area and have some of the highest numbers of acute stroke interventions. The busy centers in the country are doing 75–125 thrombectomies a year. This figure is based on personal communication and hospital discharge data. The real number is always a bit lower than we neurointerventionalists typically proclaim! If we project this number on their respective population densities we come up with an average of 10 thrombectomies for 100 000 people for these centers. If everywhere in the country we were performing 10 stroke interventions for every 100 000 people we would reach 31 800 thrombectomies annually (the range is 25 000–38 160 based on 8–12 interventions/100 000). The lower end of the range at around 25 000 may be a reasonable target—reasonable because there are centers that are achieving this for their population.
In the GRACE registry about 30% of the patients with ACS admitted within 12 h of symptom onset underwent revascularization using a PCI.6 After the initial growth in STEMI interventions from the late 1990s to the early 2000s, the rate of PCI after mid to late 2000 has been declining.7 ,8 The number of patients admitted for ACS receiving PCI varies from as low as 30% to as high as 75%,4 ,6 with the higher numbers generally reported from the earlier years and the lower numbers from more recent publications. One paper published in 2011 reported a rate of 46% PCIs for patients with STEMI.9 Assuming a 50% rate of coronary interventions and the current incidence of STEMI as previously discussed, yields 30–40 PCIs performed for 100 000 patients with STEMI or over three times the 8–12/100 000 stroke interventions performed for LVO strokes in the current high volume centers. We previously established that there are 20–50 thrombectomy-eligible LVOs per 100 000, and thus the high volume places are already treating 16–40% of their eligible LVO population.
These centers may see an increase in their volumes but are not going to shatter any growth records. The low hanging fruit in large vessel stroke treatment is to be found in centers that have a high volume of aneurysm interventions, have been the pioneers in endovascular techniques, and are at the forefront of innovative device trials—but do not do ischemic stroke. These centers do exist. I won't name any, they know who they are. These are centers with existing established infrastructure, resources, capacity, and name recognition. The only thing they have lacked so far is a desire—understandable because of an absence of high-level evidence. That has changed and now their patients, communities, hospitals, and colleagues will rightfully expect them to unveil the same classy operation for stroke that they have demonstrated for treating aneurysms. These centers are the first tier of growth for endovascular stroke therapy. The growth in practices employing one to two physicians may also occur but will be more challenging because of sharing the interventional volume with an additional physician, which brings us to the next topic.
An oversupply of workforce has been a recurring concern for our field.10 Even early on when the specialty had just found its footing, there was unease that the supply might outpace the demand.11 Over a decade ago, it was pointed out that over 80% of the US population lived within 65 miles of a neurointerventionalist and 99% within 200 miles.12 By 2012, with almost 800 neurointerventionalists, the estimated distance for access was reduced to 50 miles for 95% of the population.13 Today with over 1000 physicians in the field, access to one of us is not a concern. Since 2001, there has been an almost 25% annual increase in the number of neurointerventionalists per year with more of the growth occurring in recent years. In contrast, endovascular aneurysm volume after an early growth of 15–20% has tapered off to low single digits in more recent years. We thus have an imbalance where the supply of neurointerventionalists has not only surpassed the growth in volume but the rate of growth among the physician supply has also been completely out of sync with the volume demand. There are now more fellows being trained in more centers than ever before at a time when endovascular aneurysm treatments are reaching their peak with dwindling growth. The supply–demand gap is at its widest with no signs of shrinking. This oversupply of physicians can wreak havoc vis-à-vis endovascular stroke therapy. The centers that are currently doing a high volume of stroke interventions (8–12 per 100 000 people) are fully staffed with three to five neurointerventionalists. All these hospitals have more than 500 beds (which incidentally should be the minimum bed number required to be an interventional stroke center). Centers such as these can accommodate 400–500 of the existing workforce (based on distribution of these hospitals and current staffing). There are a handful of centers with five neurointerventionalists, a few with four, many with three, but the majority of centers doing neurointerventional procedures have one or two physicians and are working in centers with 250 to 700 beds. These one or two physician group centers may be doing 50 aneurysm treatments a year and adding another physician to cover stroke call will cut their individual aneurysm volume by a half or a third, directly affecting their revenue or relative value unit output.
In the USA, we did about 25 000 endovascular aneurysm treatments last year or 25 aneurysms per neurointerventionalist. Since the outbreak of the positive trials, I am aware of at least three situations where a one or two physician group is facing the dilemma of adding a third person because of a hospital push to cover stroke. They don't have enough volume to sustain three people but one or two physicians are not enough to sustain stroke call. Ideally, they need individuals to just cover stroke but not take away the aneurysm volume. In contrast, most centers that are designated PCI capable have three to four interventional cardiologists. With nearly 600 000 coronary interventions done annually, there is enough non-emergent volume to support four interventional cardiologists and cover STEMI call but with 25 000 aneurysm treatments and even with 30 000 thrombectomies it is going to be a challenge to accommodate the ever-growing number of neurointerventionalists. This will also increase the number of inappropriate and non-indicated neurovascular procedures and diagnostic angiography. Before the cardiology community instituted strict appropriateness criteria and registries to monitor interventional therapy, inappropriate diagnostic coronary angiography was resulting in a large number of un-indicated percutaneous interventions14 in the booming era of coronary procedures.
Another glaring difference in workforce is in its makeup. Interventional cardiology involves a homogeneous group of practitioners. We are a loose federation at best and a fractious one at worst. Whatever specialty we belong to, our motivation for advancing endovascular stroke care could be underscored by a fear of ceding ground to others. We may like to equate LVO interventions and growth to the coronary field, but other than a shared pathophysiology of acute vascular occlusion leading to critical end organ damage, the differences in practice and practitioners far overshadow the similarities. Orange may be the new black, but LVO ain't the new STEMI—at least not yet.
The accurate number of LVO cases amenable to thrombectomy is elusive and the first action required by our societies across different specialty backgrounds should be to institute and encourage documentation of all cases in a central registry. The number nonetheless is not large and reaching even 25 000 in the next few years will be a triumph; remember 8–12 thrombectomies per 100 000 people will get us there and the encouraging part is that there are centers that are currently doing that. We do need to temper some of our enthusiasm and base our growth on reality. In a free market economy there are forces that can create strong headwinds to any well thought out approach. Forces such as an oversupply of a diverse physician population that may create its own pseudo-demand, forces such as hospitals espousing “if you build it, it will come” because they see a revenue stream in this emerging ‘service line’, and, lastly, forces in the industry competing for device share in a dwindling market. I have used assumptions in the above discussion but they are based on as reliable data as are available. The growth-determining step in these assumptions is patient eligibility. Anything that makes more patients eligible like neuroprotection, cooling, or rapid transfers will allow more patients to be treated. For now, acute ischemic stroke interventions could either become resource intensive, costly, and infrequently occurring emergent procedures, managed by a heterogeneous group of physicians stumbling over each other—or we can come together, take ownership and develop a sound, viable and durable strategy based on the ground realities. This editorial was meant to give a sober overview of the field as seen by some. Hopefully a separate detailed analysis will be available later this year. In the end I will leave you with this Lewis Carroll quote “If you don't know where you're going, any road will get you there”.
Competing interests None declared.
Provenance and peer review Commissioned; internally peer reviewed.