Article Text

Download PDFPDF

Case volumes of intra-arterial and intravenous treatment of ischemic stroke in the USA
  1. J A Hirsch1,2,
  2. A J Yoo1,2,
  3. R G Nogueira2,3,
  4. L A Verduzco1,
  5. L H Schwamm3,
  6. J C Pryor2,
  7. J D Rabinov1,2,
  8. R G González1
  1. 1Department of Neuroradiology, Massachusetts General Hospital, Boston, Massachusetts, USA
  2. 2Department of Interventional Neuroradiology, Massachusetts General Hospital, Boston, Massachusetts, USA
  3. 3Department of Neurology, Massachusetts General Hospital, Boston, Massachusetts, USA
  1. Correspondence to
    Albert J Yoo, Departments of Neuroradiology and Interventional Neuroradiology, Massachusetts General Hospital, 55 Fruit Street, Gray 241, Boston, Massachusetts 02114, USA; ajyoo{at}


Background and purpose Ischemic stroke is a major cause of disability and death in the USA. Intravenous tissue plasminogen activator (t-PA) remains underutilized. With the development of newer intra-arterial reperfusion therapies, there is increased opportunity to address the more devastating large-vessel occlusions. We seek to identify the numbers of patients with stroke treated with intravenous and intra-arterial therapies, as well as to estimate the potential number of intra-arterial cases in the foreseeable future.

Methods We performed a literature search to determine case volumes of intravenous t-PA use. We extrapolated the current case volume of intra-arterial stroke therapies from the numbers of cases in which the Merci retrieval device was used. In order to estimate the potential numbers of intra-arterial stroke cases, we characterized the percentage of patients with stroke who received intra-arterial therapy at two leading stroke centers. We applied these percentages to the numbers of patients with stroke seen at the top 100, 200 and 500 stroke centers by volume.

Results The rate of intravenous t-PA use is 2.4–3.6%, resulting in 15 000–22 000 cases/year in the USA. The estimated case volume of intra-arterial therapies is 3500–7200 in 2006. Based on data from St. Luke's Brain and Stroke Institute and Massachusetts General Hospital, approximately 5–20% of patients with ischemic stroke can be treated with intra-arterial therapies. Extrapolating this to the top 500 stroke centers by volume, the potential number of intra-arterial cases in the USA is 10 400–41 500/year.

Conclusion Based on the current numbers of intra-arterial cases, our theoretical model identifies a potential for significant growth of this stroke therapy.

  • Stroke
  • Intervention
  • Economics

Statistics from

Request Permissions

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.

As the leading cause of disability and third leading cause of mortality, ischemic stroke is a major public health problem. There are 795 000 strokes per year in the USA1—and this number is projected to rise given the aging demographic of the US population. Of these, 692 000 (87%) are ischemic. Approximately 138 000–277 000 (20–40%) ischemic strokes are due to large-vessel occlusion and may thus be amenable to treatment with intra-arterial intervention.2 3 4

Reperfusion is the only proven acute intervention for ischemic stroke.5 6 7 Intravenous tissue plasminogen activator (t-PA) was cleared for acute stroke treatment in 1996. Despite having nearly 10 years to gain widespread clinical adoption, intravenous t-PA remains underutilized.8 Limited adoption of intravenous t-PA may be partially due to the restrictive 3 h time window for administration.9

Interventionalists have long been touting the promise of intra-arterial treatments for ischemic stroke because the more focused therapy allows a wider treatment window. Many patients who are contraindicated for intravenous t-PA still qualify for intra-arterial therapies. Moreover, mechanical embolectomy and localized delivery of thrombolytics allow physicians to more effectively revascularize proximal occlusions with greater clot burden—lesions that tend to be recalcitrant to intravenous t-PA.

Relatively little has been written on estimated case volumes for intravenous thrombolysis; there has been even less published on annual case volumes of intra-arterial treatment of ischemic stroke. In this article we seek to better characterize the numbers of patients who are receiving intravenous and intra-arterial treatments for ischemic stroke. Furthermore, we will endeavor to establish the potential numbers of patients who might someday qualify for these treatments.


We initially performed literature searches to determine reported case volumes of both intravenous t-PA and intra-arterial interventions.

In the case of intravenous t-PA, Kleindorfer et al have looked at the Medicare Provider Analysis and Review (MEDPAR) and the Premiere Pharmacy databases to attempt to characterize the percentage of ischemic stroke cases coded for intravenous thrombolysis through 2004.10 Until 2005, however, there was no incremental reimbursement for coding for intravenous thrombolysis, so the MEDPAR database may have underestimated the use of intravenous t-PA. We also performed a literature search of registries and regional studies that looked at the percentage of intravenous t-PA use. We multiplied the range of intravenous t-PA percentage utilizations that we found by the annual number of ischemic strokes in the USA to calculate a case volume range of intravenous t-PA use per year.

The literature is meager with respect to intra-arterial stroke therapy utilization. Until 2006, there was no specific International Classification of Diseases, 9th Revision (ICD-9) coding for mechanical embolectomy so MEDPAR utilization data have yet to become available on intra-arterial intervention. The Paul Coverdell registry has reported rates of intra-arterial intervention, but more recent case volumes of intra-arterial ischemic stroke intervention on a national or regional scale are not well characterized in the literature. To calculate a range for the annual numbers of intra-arterial interventions in the USA, we relied upon data provided to us by Concentric Medical, manufacturer of the Merci Retrieval System (Mountain View, California, USA). We obtained the number of cases in which the Merci device was used for 2005 and 2006. We then looked to the literature to determine the percentage of intra-arterial cases in which the Merci device was used in order to estimate the total volume of intra-arterial interventions.

Finally, in order to estimate the potential number of intra-arterial interventions in the foreseeable future, we determined the percentage of ischemic strokes that underwent intra-arterial intervention at Massachusetts General Hospital (MGH) for 2005. We also looked to the literature to obtain a similar figure from another highly developed stroke center (St Luke's Brain and Stroke Institute (SLBSI), Kansas City, Missouri, USA). We assumed that the intra-arterial treatment utilizations from these two hospitals could be replicated at the top 100, 200, and 500 centers in the USA. We used a 2004 database from Verispan, Inc. (Yardley, Pennsylvania, USA) which reported the number of ischemic strokes (ICD-9 codes 433.X and 434.X) by center in the USA to estimate the number of ischemic strokes presenting to the top 100, 200, and 500 centers (by volume) in the USA. We then calculated the numbers of potential intra-arterial stroke interventions if the top 100, 200, and 500 stroke centers began treating as actively as MGH and SLBSI St. Luke's Hospital.

We received approval from the MGH Institutional Review Board for the retrospective analysis of the number of stroke therapies at our hospital and the percentage receiving intra-arterial therapy.


Estimates of intravenous t-PA annual use in the USA

Kleindorfer et al searched the MEDPAR and Premiere Pharmacy databases to find estimates of intravenous t-PA utilization for years as recent as 2004.10 For 2004 Kleindorfer et al reported intravenous t-PA utilization rates of 0.83% based on the MEDPAR database and 2.4% based on the Premiere Pharmacy database. The Paul Coverdale registry showed for the 2001–2002 years, intravenous t-PA was administered to 3.6% (153/4280) of patients captured.11 12 Excluding the MEDPAR data, as it may have insufficiently captured intravenous t-PA use, the 2.4–3.6% range correlates to ∼15 000–22 000 cases/year. In the 10 years since its approval, there have only been ∼100 000 patients treated with intravenous t-PA in aggregate in the USA.13 14

Estimates of intra-arterial therapies annual use in the USA

The Paul Coverdell registry captured intra-arterial t-PA administration. In their data from four states from 2001 to 2002, they observed a 0.75% rate of intra-arterial t-PA administration (32/4280). Applying this broadly to the current 692 000 ischemic strokes annually in the USA yields an estimated annual case volume of 5200.

Concentric Medical estimates that 1400 cases employing Merci were performed in the USA in 2005, and 2300 cases were performed in 2006. At that time, Merci was the only Food and Drug Administration (FDA)-cleared intra-arterial intervention. However, additional intra-arterial interventions were being practiced, including thrombolytic administration, balloon angioplasty, guidewire mechanical disruption, and investigational interventions such as intra-arterial sonolysis.

To estimate the proportion of intra-arterial cases that employed the Merci device, we looked at published experiences in high-volume intra-arterial stroke centers. Choi et al published the endovascular experience at the Columbia University Medical Center from 2001 to 2004. In their cohort, 32% (10/31) of all intra-arterial interventions involved retrievers.15 Rymer published data showing the breakdown of all 351 patients at Mid-America Brain and Stroke Institute, St Luke's, Kansas City, Missouri (now St Luke's Brain and Stroke Institute) that received intravenous or intra-arterial intervention between 2003 and 2005. In their cohort, 65% (162/250) of all intra-arterial interventions involved the Merci retrieval device.16 The 32–65% range of Merci utilization in endovascular stroke therapy combined with the Concentric data yields a national intra-arterial case volume of 2200–4400 in 2005 and 3500–7200 in 2006 (table 1).

Table 1

Estimated numbers of intra-arterial (IA) ischemic stroke interventions in 2005 and 2006 in the USA

Projected annual use of intra-arterial therapies in the USA if the top 100, 200, 500 centers treated as frequently as major stroke centers

In 2005 at MGH, there were 570 patients discharged with a diagnosis of acute ischemic stroke; 31/570 (5.4%) of these patients underwent intra-arterial reperfusion therapy. Focusing on those patients who first presented to MGH, 264/570 (46.3%) initially presented to the emergency department and underwent stroke imaging with head CT/CTA, magnetic resonance imaging and/or magnetic resonance angiography. Of these, 67 (25.4%) were found to have a large-vessel occlusion as defined by occlusion of the intracranial internal carotid artery, proximal middle cerebral artery or basilar artery. Fourteen of the cases with large-vessel occlusion went on to receive intra-arterial stroke intervention. Therefore, the percentage of patients with acute ischemic stroke who initially presented to MGH undergoing intra-arterial therapy was 5.3% (14/264).

The group at SLBSI has published case volume data showing that among the 340 acute ischemic strokes presenting to their hospital in the first 9 months of 2004, they performed intra-arterial interventions on 20% of patients, and intravenous or intra-arterial intervention on 29% of patients.17 The lower rate of intra-arterial treatment utilization at MGH in comparison with SLBSI may be related to the greater number of treatment options available at MGH including desmoteplase (during enrollment for the DIAS-2 trial), hypertensive therapy, and hyperoxia, which compete for the treatment of the patient with acute stroke at MGH.

We used the approximately 5–20% rate of intra-arterial treatment at these two centers as a realizable estimate for the numbers of cases a given stroke center could potentially perform, and calculated the potential case volumes that could be performed if all of the top 100, 200, and 500 stroke hospitals by volume intervened as frequently as MGH or SLBSI (table 2).

Table 2

Estimated annual case numbers of intra-arterial ischemic stroke interventions at the top 100, 200, and 500 volume stroke hospitals if they intervened as frequently as groups at Massachusetts General Hospital and St. Luke's Brain and Stroke Institute, Kansas City, Missouri


The first wave of growth for intra-arterial treatment of stroke occurred around 1999 with the publication of the PROACT II study. PROACT II demonstrated that use of intra-arterial pro-urokinase in MCA stroke within the 3–6 h window was associated with improved outcomes compared with controls.6 Although pro-urokinase was never cleared by the FDA, these results have been used to justify widespread off-label use of intra-arterial thrombolytics within the 3–6 h window. Furthermore, simply conducting the study enabled 54 centers to develop their clinical pathways to allow for intra-arterial management of stroke.

A more recent wave of growth for intra-arterial treatment of stroke has resulted from the FDA clearance of the Merci Retrieval System (Concentric Medical) in August 2004. The MERCI and Multi MERCI trials evaluated the use of the device in an 8 h time window.18 19 Moreover, the FDA labeling does not specify a time window for mechanical intervention. This gives neurointerventional specialists the opportunity to offer intervention in instances where imaging might suggest viable penumbra even beyond the 8 h window.

One of the goals of this publication was to provide a first estimate of volumes of current intra-arterial stroke interventions and to speculate on what those volumes may become in the near future. Very little has been published on this topic to date. Our analysis certainly has limitations. For example, we attempted to estimate intra-arterial case mix based on data that came from only two centers and included data from years straddling regulatory clearance of Merci. We were obligated to use this methodology because of the paucity of data available on the frequent use of unapproved devices or therapies (intra-arterial thrombolytic alone, balloons, etc). Our hope is that we might stimulate others to find better ways to characterize and track rates of intra-arterial stroke intervention.

Data calculated here would seem to suggest that we are in the midst of a wave of dramatic growth of intra-arterial intervention, with as many as 7200 cases performed in 2006. This growth may be accelerated by the December 2007 FDA clearance of the Penumbra stroke device, although it is still too early to assess this impact. If this rate of growth continues, intra-arterial intervention may become as frequent as intravenous thrombolytic administration within a few years. Furthermore, based on data in table 2, the ceiling for this growth may be much higher. If the top 500 hospitals by stroke volume were to treat as frequently as the group at SLBSI, up to 41 500 interventions per year would be performed at these hospitals alone.

Cloft et al recently published a review article assessing the demand and available work force for intra-arterial stroke therapy.20 The authors sought to determine a maximum number of potential treatment cases primarily by accounting for issues that would mitigate against more widespread adoption of intra-arterial treatment. The pretext of our analysis was quite different as we sought to determine a maximum potential number of intra-arterial cases, absent normalized data, without making any assumptions about treatment delivery.

As in our case, Cloft et al are forced to extrapolate from the information that is available to project potential patients for intra-arterial stroke treatments. While their methodology is somewhat different, a number of interesting comparative points bear mentioning. The authors start with a figure of 750 000 strokes/year versus our figure of 795 000 strokes/year. Both are published and reasonable estimates. They correct by 19% for non-ischemic causes and derive therefore that in total there are 645 000 ischemic strokes/year in the USA, whereas, we methodologically correct by 13% for cases not due to ischemic causes leading to our total of 692 000 potential ischemic strokes/year. If anything, these numbers are noteworthy for their similarity and not their difference. The authors of that paper make many erudite points that we find ourselves in agreement with including the growing number of Society for NeuroInterventional Surgery members treating stroke and the proliferation of neuroendovascular fellows. In addition, they point to the need for a comprehensive approach to stroke delivery including improved access to advanced imaging, rather than a “top down” approach beginning with intra-arterial therapy. Finally, Cloft et al describe the possibility of 100 patients/year being treated at the top 200 comprehensive stroke centers and note that that leads to the possibility of 20 000 intra-arterial cases/year as a ceiling on the number of potential cases. We provide an alternate methodology of assessing potential intra-arterial treatment numbers using our own and other published experiences and relate that experience to potential numbers of cases if that approach were duplicated in 100, 200 and 500 medical centers. In fact, the range we provide for potential numbers of intra-arterial cases in the top 200 stroke hospitals was 5400–21 700 intra-arterial cases/year. We simply do not agree that this represents the ceiling on the number of potentially treatable patients with ischemic stroke treated with intra-arterial means.

There are some recent events that may help to further drive the current growth of intra-arterial intervention. In October 2006, Centers for Medicare & Medicaid Services put in place separate coding for hospital reimbursement for patients who are medically managed, those who receive thrombolytics (intravenous and/or intra-arterial), and those who have mechanical thrombectomy. National averages for typical Medicare Severity Diagnosis Related Groups (MS-DRGs) for these different treatment options are provided in table 3. The higher paying MS-DRGs for thrombolytic and mechanical intervention will help centers dedicate resources to more aggressive management of stroke, including 24 h access to intra-arterial intervention.

Table 3

Average Medicare reimbursement for different ischemic stroke treatment options

Trials such as MR RESCUE, IMS III, and others may be able to bolster the level of evidence for intervention thereby encouraging non-interventional centers to transfer their large-vessel strokes to centers with access to intra-arterial therapies, thereby raising the numbers of patients finding their way to the top stroke centers.

More widespread use of advanced imaging modalities such as CT angiography, CT perfusion, and diffusion/perfusion magnetic resonance imaging in the hyperacute setting has the potential to revolutionize ischemic stroke treatment. CT angiography can accurately identify large-vessel occlusion, which is critical in deciding which patients are suitable for endovascular therapy. Moreover, imaging selection based on patient physiology may extend therapy to the greater than 20% of strokes that have an undetermined time of onset, as well as to the greater than 40% of strokes that present beyond 6 h from onset (fig 1).9 11 21

Figure 1

Percentage of patients arriving within various time windows in the Paul Coverdell registry, University of Cincinnati experience (Kleindorfer), and Buffalo Metropolitan area and Erie County experience (Qureshi).

For this growth in intra-arterial stroke treatment to occur, there needs to be a dramatic increase in the number of neurointerventionalists. Percutaneous coronary intervention (PCI) for myocardial infarction serves as a good model for the number of interventionalists required to provide round-the-clock coverage for an acute disease state. Each year in the USA, approximately 6500 interventional cardiologists perform approximately one million PCIs, approximately one-quarter of which are for acute myocardial infarction.22 23 Current data indicate that there are potentially 138 000–277 000 large-vessel strokes. While many of these patients are not appropriate candidates for intra-arterial therapy for a variety of reasons, the potential treatment pool serves as both a challenge and an opportunity for the neurointerventional community. As the group most trained and capable of intervening in the cerebral vasculature, neurointerventional specialists must continue to take the lead in addressing this potential patient need.

The authors acknowledge that many things need to occur prior to the wholesale increase in acute intervention alluded to in this article. There remain many barriers to the improvement of the care of patients with stroke well before the interventional suite. These include education of the patient and society of the need for early introduction to the healthcare system—for example, the Brain Attack platform that has been introduced in recent years. Additionally, emergency medical services and emergency room physicians require continued education about the need to triage patients to primary stroke centers. Imaging paradigms that are very helpful in the acute management of the patient with stroke are not available in every location, and there remains no consensus on the best imaging approach.


The purpose of this article is to assess the potential growth in the number of intra-arterial stroke interventions in the not too distant future. We view this as similar to what occurred in the management of acute myocardial infarction when percutaneous transluminal coronary angioplasty and stenting transformed that field 15–20 years ago. There has been a perfect storm of events that will enable intra-arterial treatment of stroke to grow for the foreseeable future: FDA clearance of tools for revascularization, continued improvement of diagnostic imaging, and improved reimbursement.

Our theoretical models suggest that ischemic stroke is a field that has the potential to be substantial in size. As a point of reference, it could impact 5–10 times the number of patients that are endovascularly managed for hemorrhagic stroke secondary to ruptured cerebral aneurysms. Neurointerventional specialists remain best positioned to lead the continued growth of appropriate intra-arterial treatments for patients with stroke.


The authors acknowledge Neil Barman, for editorial contributions.



  • See Editorial, p 8

  • Competing interests JH: Merci Registry Steering Committee (no financial compensation); RN: Concentric Medical, Inc. (Physician Advisory Board, Consultant, Lecturer - modest, no stocks), ev3 (Physician Advisory Board, Consultant, Lecturer - modest, no stocks), Coaxia (Physician Advisory Board, Consultant, Lecturer - modest, no stocks); LS: Coaxia Inc (medical advisory board, compensated consultant for stroke device trial design), RTI Health Inc (compensated consultant, economic analysis of MRI-selected patients for thrombolysis; JP: ev3 Inc (procedural proctor, lecturer, advisory board - no financial compensation); RG: Bayer, GE (consultant - no financial compensation).

  • Funding AY was the 2007 recipient of the Neuroradiology Education and Research Foundation/Boston Scientific Fellowship in Cerebrovascular Disease Research. LAV was a 2007 recipient of the Howard Hughes Medical Institute Research Training Fellowships for Medical Students.

  • Ethics approval We received approval from the MGH Institutional Review Board for the retrospective analysis of the number of stroke therapies at our hospital and the percentage receiving intra-arterial therapy.

Linked Articles