TwitterRSSFacebookDirectors Blog
  Disorders A - Z:   A    B   C    D    E    F    G    H    I    J    K    L    M    N    O    P    Q    R    S    T    U    V    W    X    Y    Z

You Are Here: Home  »  About NINDS  »  Advisory and Peer Review Groups  »  Stroke Progress Review Group  » 

Skip secondary menu

Report of the Stroke Progress Review Group - September 2006

Additional information about Stroke PRG

Update: Report of the Stroke Progress Review Group, January 2012

Table of Contents

Executive Summary

Acute Stroke Treatment
Biology of Repair
Clinical Trials
Cerebral Vascular Biology and Neurovascular Unit
Stroke Epidemiology and Risk Factors
Health Services Implementation
CNS Hemorrhage/ICH
Endothelium and Hemostasis
Neurocererovascular Degeneration
Neurovascular Protection
OMICs - Genomics, Proteomics, Metabolomics and Bioinformatics
Prevention of First and Recurrent Stroke
Recovery and Rehabilitation Working Group
Vascular Cognitive Impairment

SPRG Meeting Participants
SPRG Meeting Working Group Members


Co-Chairs: Helmi Lutsep, Patrick Lyden, Jeffrey Saver

Members: Kyra Becker, Jeff Frank, Kama Guluma, Clarke Haley, Michael Hill, Argye Hillis, Rebecca Ichord, Reza Jahan, Tom Kwiatkowski, Brett Meyer, Larry Weshsler

NINDS Liaison: Scott Janis

1. Seminal scientific advances made since 2001. Please indicate if they are linked (or not) to SPRG recommendations.

A. Reperfusion: The restoration of blood flow into ischemic brain remains the only successful treatment for stroke, so far. This therapeutic area was the first priority in the original PRG report, and considerable progress occurred since the Implementation report.

To review the original report: Priority 1: Reperfusion: therapeutic agents that open blood vessels in more patients and that do so better, faster, and more safely, are greatly needed.

Since identification of this priority for acute ischemic stroke treatment in 2002, progress in reperfusion therapies has advanced along several fronts. NINDS funded a dose escalation safety study of tenecteplase that was completed in 2003. Promising results from that experience resulted in funding of a larger Phase 2B study comparing 3 different doses of tenecteplase to each other and to standard treatment with alteplase. Another long-acting and more fibrin specific agent, desmoteplase, underwent preliminary safety testing using MRI-based selection criteria to allow treatment within 3 and 9 hours from stroke onset. Further study of this compound is also now in progress. Early studies examining plasmin, microplasmin, and alfimiprase are underway. A trial of acute treatment with a glycoprotein 2b/3a inhibitor, abciximab, within 5 hours of stroke onset had to be stopped early because of an unfavorable risk/benefit ratio.

NINDS funded trials sought to enhance the efficacy and/or safety of standard fibrinolysis. Phase 2 studies examining combination of glycoprotein 2b/3a inhibitors (CLEAR, ROSIE), a prothrombin inhibitor (argatroban), and hypothermia with standard alteplase or reteplase are in progress or nearing completion. A phase 2 trial of high-frequency transcranial Doppler ultrasound in combination with alteplase (CLOTBUST) showed promising results, and a phase 3 trial has been funded. A trial examining non-focused, low frequency transcranial ultrasound in combination with alteplase (TRUMBI) was stopped early because of excessive intracranial bleeding complications. A phase 3 trial comparing standard dose intravenous alteplase within 3 hours of onset to low dose intravenous alteplase administered within 3 hours followed by intra-arterial recanalization therapies (IA alteplase, IA ultrasound catheter delivered alteplase, or the Merci Retriever), if necessary, within 6 hours has been recently funded by NINDS, after a phase 2 study showed promising results (IMS 2). An industry sponsored phase 3 trial of Ancrod, administered within 6 hours of symptom onset is also in progress.

Progress has been made in extending the time window for reperfusion of imaging to select for treatment late presenting patients still harboring salvageable penumbra. A NINDS-funded trial of MRI screening of patients treated with alteplase in the 3-6 hour time period found that 40-50% of patients exhibit a small infarct, large penumbra pattern and appeared to benefit from therapy (DEFUSE). Imaging selection of patients was also used for the desmetoplase trials (3-9 hour window) and for the ongoing glycoprotein 2b/3a inhibitor plus reteplase therapy trial (time window up to 24 hours) (ROSIE).

Experience continues to accumulate with a number of mechanical devices to disrupt/remove occlusive cerebrovascular thrombi. Techniques explored include transcatheter-applied local ultrasound, laser, or suction devices using the Venturi principle. Various clot retrieval devices including snares, "corkscrews," and others show promise in the removal of occluding thrombus. The U.S. Food and Drug Administration recently approved the MERCI retrieval device specifically for removal of intracranial cerebrovascular clots; however, the agency acknowledged that the efficacy of this approach for stroke treatment has not been proven. A NINDS-funded randomized clinical trial of the MERCI device (MR RESCUE) is under way.

In summary, since the last review, NINDS funded initiatives successfully implemented priority 1, with several new agents and treatment strategies in late-stage development. Further, a few agents that looked promising were found to cause excessive bleeding, allowing investigators not only to winnow the field of candidate therapeutics, but also to gain further insights into the mechanisms of thrombolytic-related hemorrhages.

B. Neuroprotection: The original report listed cytoprotection as Priority 3: Priority 3: Clinical trials and establishing the utility of cytoprotection: a shift is needed from single agent trials to combination trials.

Although a number of clinical trials have reported out since the Implementation report, including one highly promising trial (SAINT 1) there remains no approved effective neuroprotection. Further, NINDS funded initiatives yielded a handful of combination trials, primarily from the SPOTRIAS Network, examining multiple neuroprotective agents, or thrombolysis combined with neuroprotection.

The SAINT I Study, a Phase III prospective randomized controlled trial assessing the benefit of NXY-059 (a putative free radical scavenger manufactured by Astra-Zeneca), showed that neuroprotection remains a viable strategy for the treatment of stroke. In SAINT I, patients who received NXY-059 within 6 hours of the onset of ischemic stroke were more likely to improve than patients who received placebo (odds ratio, 1.20; 95% CI, 1.01 to 1.42). Moreover, among patients who received IV tPA, those treated with NXY-059 were significantly less likely to suffer a hemorrhagic transformation of their infarct, suggesting a vascular protective property to the drug as well. SAINT II, a confirmatory study of NXY-059, failed to confirm these promising findings. .

Cytidyl diphosphocholine, or citicoline, is a naturally occurring nucleoside essential for the formation of phosphatidyl-choline and subsequently for maintenance of membranes. The North American Study of citicoline in stroke failed to show a benefit in the primary end point, but a pooled analysis of patients treated with this drug in multiple trials suggests that it may improve stroke outcome. A large international trial, International Citicoline Trial on Acute Stroke (ICTUS) is currently enrolling patients enrolled within 24 hours of stroke onset. Citicoline was approved in several countries for acute stroke.

The SAINT studies and the citicoline studies are sponsored by the Pharmaceutical Industry. The NIH also funded several ongoing trials of putative neuroprotection agents. Based on animal studies and early pilot studies in humans, high dose albumin soon after stroke onset appeared to be neuroprotective. The results of a Phase II study of Albumin in Acute Stroke (ALIAS) were recently published. In this study, patients treated with 2.05 g/kg of 25% human albumin within 16 hours of stroke onset were significantly more likely to have a better outcome than patients treated with lower doses of human albumin or similar patients in the NINDS tPA study. Based on the findings of this study, a Phase III study of high dose human albumin therapy for patients who can be treated within 5 hours of acute ischemic stroke is underway.

Many neuroprotective agents can safely be given to patients with both ischemic and hemorrhagic stroke. Accordingly, they could be started by paramedics in the ambulance prior to hospital arrival, speeding treatment start. Based on extensive experience with magnesium as an experimental neuroprotective agent in various types of brain injury, it appears to be a logical drug to test in acute stroke. A NINDS-funded pilot study of magnesium administration in field was performed demonstrating the feasibility and safety of magnesium administration in the field, with treatment start accelerated by 2 hours; a NINDS-funded Phase III study, the Field Administration of Stroke Therapy - Magnesium (FAST-Mag) Study, is now recruiting patients with presumed stroke who can receive treatment in the field within 2 hours of symptom onset.{ .

Hypothermia is a promising treatment for stroke given the abundance of animal data on its robust neuroprotective properties and the demonstrated benefits on neurological outcome in patients with cardiac arrest. Feasibility studies of endovascular cooling have been performed. The role of hypothermia in the treatment of acute stroke and it's ability to prolong the time window for intravenous thrombolysis is being evaluated in the Intravascular Cooling in the Treatment of Stroke - Longer tPA window trial, ICTuS-L. ICTuS-L is supported through the Specialized Program of Translational Research in Acute Stroke (SPOTRIAS) initiative of NINDS. Patients treated with rt-PA under 3 hours from symptom onset are randomized to cooling or no cooling. Patients presenting within 3 to 6 hours of symptom onset are randomized to cooling or no cooling, with or without rt-PA. The aim of this safety/feasibility study is to establish safety of the combination cooling plus rt-PA prior to a larger efficacy trial. Animal data suggest that the combination of caffeine and ethanol (caffeinol) are neuroprotective; a pilot study showed that this combination was safe in patients with stroke. A study of combinatorial therapy with caffeinol (treatment within 240 minutes) and hypothermia (treatment within 300 minutes) is currently recruiting patients. This study is also funded through the SPOTRIAS program.

In addition to the successes listed previously, seemingly well designed trials conducted since 2002 generated negative results. Repinotan is a serotonin agonist that showed efficacy up to 5 hrs after symptom onset in rodent studies but was not assessed in larger animal models. A phase 3 trial in which the drug was administered within 4.5 hrs of stroke onset and plasma levels were optimized did not show efficacy. ONO-2506 is an astrocyte activation inhibitor that was administered within 6 hrs of stroke onset to patients with cortical strokes. A futility analysis done in May 2005 led to discontinuation of the trial in the United States. Subsequent sub-group analysis identified a potential benefit, and further studies may be organized. YM872 is an AMPA receptor antagonist that was evaluated in two recent phase 2 studies. One included patients within 6 hrs of symptom onset that had moderately severe strokes and assessed lesion size by MRI at 28 days as well as clinical outcome at 90 days. Another study included only those patients that had also received intravenous tPA within 3 hrs of symptom onset. Both trials were stopped in January 2003 for futility and there are no plans to develop this drug further.

SUN N4057 or piclozotan (Daiichi Asubio Pharmaceuticals, Inc.) is a serotonin agonist currently enrolling patients in a phase 2b study. The patients are required to be within 9 hrs of symptom onset (50% within 6 hrs and 50% within 6-9 hrs) and to have a measurable penumbra on magnetic resonance imaging (MRI). This compound is substantially similar to repinotan in mechanism. A neuroprotectant that is in an earlier phase of investigation is TS-011 (Taisho Pharmaceutical Co. Ltd.). This agent blocks the synthesis of 20-hydroxyeicosatetraenoic acid (20-HETE), a potent vasoconstrictor that may contribute to ischemic injury. TS-011 has been shown to reduce infarct size in rats and monkeys and is currently in a phase 1b dosage finding trial in stroke patients. As part of the SPOTRIAS program at Columbia University, a study called Neuroprotection with Statin Therapy for Acute Recovery Trial (Neu-START) is performing a dosage finding trial using lovastatin. Patients are treated within 12 hrs of ischemic stroke onset. A pilot study has been performed evaluating normobaric oxygen in the form of high-flow oxygen therapy delivered via facemask. Patients were within 12 hrs of symptom onset and also had mismatch on MRI. A larger study of normobaric oxygen is supported by the SPOTRIAS program at Massachusetts General Hospital.

In summary, combinations of neuroprotectants, or neuroprotectants plus reperfusion therapy, are likely to have effects above those of each agent alone. Combination therapies elicit great interest for future studies.

C. Stroke centers: In the original report, the need for certified stroke centers was identified as a critical needed resource. Since that report, the Joint Commission for the Accreditation of Health care Organizations (JCAHO) rolled out a national program for Primary Stroke Center Certification following guidelines issued by the Brain Attack Coalition. In September 2006 the JCAHO website listed 252 certified primary stroke centers. In addition, several states, including New York and Massachusetts, have implemented state-run stroke center designation programs using criteria similar to the JCAHO system. Altogether, as of May 2006, seven states with 28% of the American populace had implemented or were in the process of implementing designated Stroke Center systems under which Emergency Medical Services bring key stroke patients directly hospitals that have met minimal standards for acute stroke care. The cadre of certified primary stroke centers provides a ready supply of clinical trial sites; the availability of this trial site pool already makes an impact on the organization of NINDS and pharmaceutical sponsored trials. The Brain Attack Coalition has also released a set of Comprehensive Stroke Center recommendations. A certification program for Comprehensive Stroke Center designation is under review at JCAHO, where an American Heart Association (AHA) committee has been formed to advise JCAHO. Activities include a survey to address how Comprehensive Centers would impact the functioning of primary stroke centers. Further, the AHA Stroke Council has commissioned a writing group to review proposed performance indicators for comprehensive stroke centers.

In addition to identifying the need for stroke center designation, the PRG report mentioned the need for such centers to organize into clinical research consortia or networks. Three such networks (NETT, SPOTRIAS, and the Neurology Clinical Trials Consortium) are described elsewhere in detail.

D. Pediatric stroke: Recent data confirm stroke is a substantial problem in pediatric age group. The safety and efficacy of thrombolysis in children is unknown, and several ad hoc registries are in progress. Data on safety and efficacy of lytics and other therapies in pediatric patients is needed.

E. Emergency Department Investigators and Protocols: The Progress Review Group identified the need to integrate emergency department investigators in collaborative stroke research initiatives a new priority, and considerable gains have been made since 2001. Similar to patients with acute cardiac conditions, patients with acute stroke are initially evaluated and treated by Emergency Medicine personnel, including Emergency Medicine physicians and pre-hospital personnel (EMT's and paramedics). As the outcome of stroke is often determined within a very brief time-frame following the onset of symptoms, interventions in the emergency department, or even prior to emergency department arrival, may provide meaningful benefit. Recent progresses in the development of emergency department investigators as part of large collaborative research initiatives include the following programs:

1) Neurological Emergencies Treatment Trials (NETT) Network: Clinical Coordinating Center (RFA-NS-06-002), Statistical Center (RFA-NS-06-008) and Hubs (RFA-NS-06-009)

The NINDS NETT network creates an infrastructure to promote and conduct clinical trials that will provide new and effective treatments for neurologic emergencies. The key elements of this structure include an interdisciplinary team of emergency physicians, neurologists, neurosurgeons and biostatisticians with specific expertise in the conduct of emergency clinical trials, a hub-and-spoke network design that acknowledges the importance of enrolling patients in both large academic centers and community hospital emergency departments, and a long term network commitment to advancing the national research infrastructure for the interdisciplinary study of emergent illness. In addition to the Network Coordinating Center and the Statistical Data Management Center, up to eleven clinical hub sites will be funded as part of the NETT Network. Each hub site will coordinate activities at 2-10 emergency departments or "spokes". The goals of the NETT Network are: 1) to facilitate high-quality clinical trials in several different types of emergency neurological disorders afflicting adults or children; 2) to encourage collaboration between emergency medicine physicians and neurological disease specialists in trial design and execution; and 3) to facilitate the implementation of new therapies into clinical practice. Once established, the NETT Network is intended to serve as a resource for investigator-initiated clinical research in acute neurological disorders. Following the initial trials funded through this initiative during the development phase, subsequent trials will be funded through traditional NINDS peer-review mechanisms.

2) Field Administration of Stroke Therapy - Magnesium (FAST-MAG) Trial (RO1NS4436). Clinical Coordinating Center: UCLA School of Medicine. Principal Investigator: Jeffery Saver, MD

The FAST-MAG Phase 3 Trial is a multi-center, randomized, placebo-controlled, double-blind, parallel group trial of intravenous magnesium sulfate initiated by paramedics in the field within 2 hours of symptom onset in patients with acute stroke. The primary objective of the study is to evaluate the efficacy and safety of field-initiated magnesium sulfate in improving the long-term functional outcome of patients with acute stroke. The study includes patients within the service area of the Los Angeles County Emergency Medical Services Agency, whose component systems provide pre-hospital care to a population of 9.8 million. Patients with acute stroke are identified in the field by licensed paramedics who have received training in basic and advanced cardiac life support, stroke recognition, and specific procedures relevant to the proposed study. Physician-investigators approve each patient for study entry after cellular phone contact with paramedics. By phone, physician-investigators also elicit informed consent to participate in the study-from patients when the subject is competent and from on-scene legally authorized persons when the subject is not competent. The central aim of this study is to demonstrate that paramedic initiation of the neuroprotective agent magnesium sulfate in the field is an efficacious and safe treatment for acute stroke. Successful conduct of the trial will serve as a pivotal test of the promising neuroprotective agent magnesium sulfate in acute stroke, and will also demonstrate for the first time that field enrollment and treatment of acute stroke patients is a practical and feasible strategy for phase 3 stroke trials, permitting enrollment of greater numbers of patients in hyperacute time windows. An initial pilot trial which enrolled 20 patients was completed on 2002. The current phase III trial has been enrolling patients since January 2005.

3) The Resuscitation Outcomes Consortium (RFA-HL-04-001)
Data Coordinating Center: University of Washington, Seattle, WA
Regional Clinical Centers:

  • University of Alabama at Birmingham in Birmingham, AL
  • University of Texas Southwestern Medical Center in Dallas, TX
  • University of Iowa in Iowa City, IA
  • University of Pittsburgh in Pittsburgh, PA
  • Oregon Health & Science University in Portland, OR
  • University of Ottawa, Ontario and St. Paul's Hospital, University of British Columbia, British Columbia, Canada
  • University of California at San Diego in San Diego, CA
  • University of Washington, Seattle, WA
  • University of Toronto in Toronto, Ontario, Canada

The Resuscitation Outcomes Consortium (ROC) was developed in order to create the network of 10 Regional Clinical Centers (RCCs) above, through which collaborative clinical research in the arena of resuscitation in out-of-hospital cardiac arrest and traumatic brain injury can be efficiently performed by enrolling large numbers of patients (15,000 patients over a 3-year period). The research foci planned for the consortium include the development of new drug therapies, fluid resuscitation in traumatic brain injury, immunomodulation of secondary injury, neuroprotection, bleeding control, alternative cardiopulmonary resuscitation approaches, and new ventilation and oxygenation strategies. The ROC is in the very initial stages of implementation, and the first trials currently slated for initiation include an investigation of hypertonic saline and dextran and neurological outcome in severe traumatic brain injury, and an investigation of early defibrillation and the use of a one-way valve in ventilation in outcome after cardiac arrest. While focal cerebral ischemia (ischemic stroke) is not a defined research focus per se, the mission and activities of the consortium will bring mechanisms and therapies affecting global cerebral ischemia into the forefront of Emergency Medicine research.

F. Telestroke: In the original report, Telemedicine was identified as a needed resource. Since then, the application of telemedicine in acute stroke (so-called Telestroke) has advanced considerably. Over the past 10 years, telemedicine in stroke has evolved from an idealistic dream to a somewhat commonplace reality. The initial question of teleconsultation's feasibility in acute stroke management has been addressed, now allowing for more complex questions of efficacy, safety and future possibilities. The majority of acute stroke victims do not reach the acute care facility within the 3 hour time window mandated for FDA approved rt-PA use. Lack of, or delay in, acute treatment is associated with poorer stroke outcomes. Though nationally the rate of rt-PA treatment varies between 1- 5%, even in established stroke treatment facilities, the rate may only reach 3- 8.5% as shown in the Coverdell Registry. One driving factor resulting in low thrombolytic rates is the lack of specialty expertise available in healthcare shortage areas or in neurologically underserved facilities. In addition to delays in patient arrival and absence of specialty practitioners, there may be hospital delays, due to unfamiliarity with therapies, procedures and protocols, which result in a decreased treatment rate.

The current models of maximizing thrombolytic evaluations have enabled more thrombolytic treatments, but can not be sustained indefinitely nor can they be applied to diffuse geographic regions. Each of these techniques emphasizes an opposite side of the 'hub & spoke' model. The Houston "Mecca Model" relies on patients being quickly transported to a hub facility, while the San Diego "Commando Model" relies on the practitioner being able to quickly drive to the spoke. Telemedicine 'hub & spoke' techniques facilitate 2 way exchanges of patient information and practitioner expertise irrespective of distance.

Telemedicine has been applied to numerous fields of medicine with large degrees of success. Feasibility has been shown in areas such as trauma, psychiatry, cardiology, critical care, neurosurgery, and others. Telemedicine's use has exponentially increased, though clinical efficacy trials have lagged far behind. Roine et. al. published a review noting only 6 randomized, controlled clinical trials of telemedicine out of > 1000 articles assessed. A perception of 'assumed efficacy' and 'minimal harm' likely drive telemedicine's increasingly popular use. In 1999, Levine and Gorman explained both the promises and limitations of telemedicine for use in stroke. Limitations included high cost, lack of reimbursement, lack of clinical standards, scheduling difficulties and time restrictions. Many of these areas have since been addressed though some require further support and championing before telemedicine can be universally applied.

Since 2001, general system designs evolved to include 2-way audio and video capabilities enabling real time face to face conversations with both patient and requesting practitioner. Further requirements include high resolution (400 x 300) capabilities, full screen video abilities, full frame video rate (30 frames per second), synced audio with video, echo cancellation, high data compression and bandwidth, HIPAA compliant security measures, DICOM radiology CT image access and viewing software, and the ability to access a system with minimal, or no delay. To date, technical requirements and clinical protocols have not been standardized. There are currently numerous and varied telemedicine systems in use ranging from 'off the shelf' ISDN bound point to point teleconferencing software to highly complex, site independent, web based access or Internet based systems.

Though there are clearly areas for future development, much of telemedicine's promise has already been realized. Stoke telemedicine's reliability is now well documented both in the United States (Shafqat, Schwamm, Hess, LaMonte, and Meyer) as well as abroad (Wiborg, Handshu, and Audebert). Assessing for clinical deficit using the NIHSS is feasible and reliable with only minimal bedside assistance. Though felt to only be feasible 7 years ago, acute clinical consultation use is now routine in many areas. The TEMPiS- telemedical pilot project for integrative stroke care noted the ability to determine 250 different diagnoses in a 12 month period, with many crosschecked for accuracy, showing that telemedicine can be used for assessing numerous neurologic diagnoses. Another clinical use of telemedicine is in determining telemedicine's ability to more appropriately (either increase or decrease) hospital transfers for neurologic reasons. Investigators in Massachusetts showed that transfers for neurologic reasons were decreased when using telemedicine for stroke. Having telemedicine capabilities may also streamline Life-flight involvement for more rapid or increased hospital transfers, thus facilitating an improved "drip & ship" system.

G. Financial Barriers to Acute Stroke Care: In the original report, reimbursement was listed as a barrier to acute therapy. NINDS, in collaboration with several voluntary health organizations and professional societies, helped draw attention to this considerable issue. In response, DHHS promulgated new regulations, a new DRG, and a new ICD-9 Procedure Code. Until recently DRG assignments for stroke care included only 2 DRGs, DRG 14 (Intracranial Hemorrhage or Cerebral Infarction) and DRG 15 (Nonspecific CVA and Precerebral Occlusion Without Infarction). In 1998 a new code 99.10 was created (Injection or infusion of thrombolytic agent) to identify cases in which a thrombolytic agent was administered. Although this code was to be used for patients with stroke receiving IV tPA, it was not associated with any additional reimbursement.

In 2004, CMS met with representatives from major stroke centers to discuss changes to stroke DRG 14 and 15 related to thrombolytic therapy. They argued that IV tPA administration was a marker for increased stroke severity and that hospital care for such patients was significantly more costly. This increased cost was due to greater utilization of intensive care units, higher diagnostic imaging costs and increased laboratory and pharmacy costs. In addition to higher costs, IV tPA use resulted in improved outcomes for stroke patients. They recommended modification of the stroke DRGs to compensate hospitals for the higher costs associated with care of patients receiving thrombolytic treatment. Higher reimbursements for these patients would encourage hospitals to establish stroke centers to effectively treat acute stroke patients. In response to these recommendations and subsequent supportive public comments, CMS analyzed costs in the MedPAR data regarding patients in DRGs 14 and 15 with and without code 99.10. Charges for patients with code 99.10 in DRG 14 were approximately $16,000 higher than those without this code. As a result, in August 2005 CMS created DRG 559 (Acute Ischemic Stroke with the Use of a Thrombolytic Agent). This became effective in October 2005 and provided an increase in hospital payment from $4,000 to $6000 for DRG 14 to a base rate of $11,578 for DRG 559.

The Merci clot retriever was recently approved by the FDA for removal of thrombus from intracranial arteries in patients with acute stroke. In 2006, CMS in 2006 approved a new ICD-9 Procedure Code, 39.74 - Endovascular removal of obstruction from head and neck vessel(s), to be used in conjunction primarily with the craniotomy DRGs 001 and 543 for ischemic stroke patients treated with the Merci Retriever. The average reimbursement of these DRGs ($17,700 for DRG 1, $22,800 for DRG 543) reflects the higher costs for reasons similar to thrombolytic patients but also the additional expense of angiography suites, catheters and embolectomy devices.

H. Diversity Issues: The previous report identified diversity issues as a potential confound in the interpretation of ongoing clinical trials. Since then, two re-analyses of randomized trials indicated a gender-by-thrombolysis interaction in favor of women: women receive greater benefit from thrombolysis when compared to men for both IV therapy (Kent et al) and IA therapy (Hill et al). Several case series have suggested possible explanations including increased recanalization but this is not uniformly proven. One group, for example, has suggested that clot length is longer in men than women in middle cerebral artery occlusions and that clot volume tends to be larger (Buck et al). Several series also show that women who are untreated have worse outcomes compared to men who are untreated; thrombolysis neutralizes this gender bias resulting in the statistical interaction.

Certain neuroprotective agents may also have gender dependent effects. A selective kappa opioid receptor agonist has provided ischemic neuroprotection in male but not female rats (Chen CH et al). Thus far, clinical neuroprotectant trials in acute stroke have not been powered to look for efficacy differences between men and women. At least one agent, however, the free radical scavenger tirilazad mesylate, has shown poorer functional outcomes in women than in men, likely due to differences in drug metabolism (Tirilazad International Steering Committee).

There are a number of socioeconomic studies which have shown race, socioeconomic status as important prognostic factors in stroke risk, stroke recurrence and access to treatment. Fewer studies address the effects of race on treatment outcome. One group used data from a discharge database of 2594 patients treated with thrombolysis for acute ischemic stroke to assess factors associated with in-hospital mortality. In this patient population, multivariate logistic regression showed that Asian/Pacific Islander race as well as advanced age, congestive heart failure, and atrial fibrillation/flutter independently predicted in-hospital mortality after thrombolysis. No study to date has shown a differential treatment effect by race. While this remains an important social issue, it remains unclear whether differences in biology will show truly differing treatment effects in acute stroke. Further research into race/ethnicity as a predictor of treatment outcome is required.

I. Stroke Mechanisms: Although not specifically singled out in the previous report, in the past 10 years it has become clear that acute stroke trials that include too broad a range of stroke sub-types are underpowered to identify treatment effects in any single sub-group. Therefore, the proper identification of stroke mechanisms assumed larger importance. Arterial ischemic stroke (AIS) can result from atherosclerotic occlusion or critical stenosis of a medium to large artery, embolus from a clot in the heart, embolus from an atherosclerotic plaque in a proximal vessel, paradoxical embolus from a right to left shunt in the heart (or pulmonary AVM), carotid or vertebral dissection, lipohylinosis or microatherosclerosis of small cerebral vessels, or other sources. It is plausible that these different stroke mechanisms would respond differentially to various stroke therapies. However, most randomized controlled trials of treatment for AIS included patients with all of these etiologies. A few limited enrollment to patients with large cortical stroke or lacunar stroke, which are relatively easy to differentiate by clinical presentation and/or conventional MRI. The few trials that have shown efficacy in clinical outcome have included overcome this heterogeneity by enrolling huge numbers of patients to compensate for small mean effect sizes (IST, CAST), testing an agent with a substantial treatment effect (NINDS Study), or by constraining the trial population to a tightly defined subtype (PROACT II). It would be more cost-effective and efficient to identify the stroke subtypes that would be most likely to respond to treatment, or that would be likely to show larger effect sizes.

Several ongoing studies have proposed to develop MR imaging methods to differentiate stroke subtypes that might differentially respond to various interventions. For example, a NINDS funded study, DEFUSE, identified subgroups of patients with AIS, defined by diffusion and perfusion MRI profiles, who were likely to have the best response to intravenous thrombolysis administered 3-6 hours after symptom onset. The NINDS SPOTRIAS Program MR RESCUE trial is identifying MR signatures that identify patients who benefit from endovascular thrombectomy in the 3-8 hour time window. Another NINDS funded study will employ sophisticated MRI techniques to determine quantitative thresholds of specific hemodynamic measures to differentiate critical hypoperfusion from benign oligemia, that might be useful for identifying subgroups of patients who will respond to specific treatments. Similarly, an NINDS funded clinical trial will determine if specific hemodynamic measures of diffusion-perfusion mismatch are useful for determining which patients are most likely to respond to pharmacologically-induced blood pressure elevation within 12 hours of AIS onset, based on earlier published pilot studies by Rordorf, et al. and by Hillis, et al. suggesting that patients with large vessel arterial stenosis may be most likely to show improved function and/or improved perfusion with this intervention. An NINDS funded trial will validate a novel MRI pulse sequence for quantitative CBF in canines and patients with AIS and chronic ischemia. The study will test the hypothesis that qCBF can stratify risks of thrombolysis in patients who would not normally be eligible for thrombolysis. Another trial will seek to improve PWI analysis to identify "tissue at risk". A Phase II clinical trial, ROSIE - ReoPro Retavase Reperfusion of Stroke Safety study - Imaging Evaluation uses MRI criteria, as well as standard criteria, to select patients for a combination of thrombolysis (reteplase) and the platelet GP IIbIIIa antibody, abciximab, up to 24 hours from stroke onset. This study will not, however, determine whether patients who meet MRI criteria will show better response than patients who do not meet MRI criteria. Another NINDS funded trial will develop novel MRI techniques (sodium MRI and pH MRI) to identify patients who might be candidates for reperfusion therapies beyond 3 hours. In addition to MR based studies, an ongoing NINDS-funded study will test the hypothesis that early computed tomography angiography (CTA) can identify patients with specific stroke mechanisms that will respond to specific mechanism- directed therapies. The usefulness of CTA in the early AIS will be evaluated by measuring its positive predictive value, sensitivity, specificity, and reliability, and will compare CTA to current methods of clinical stroke diagnosis in a cohort of 150 patients within 24 hours of onset of AIS. The trial will determine whether CTA is a useful tool for identification of stroke etiology and selection of patients with specific stroke mechanisms for certain mechanism-directed therapies.


Grant NumberPi NameProject Title
5K24NS044848-03Albers, GregoryNew MRI techniques for cerebrovascular diseases
1R01NS049395-01A2Carroll, TimothyImproved Measurement of Celebral Perfusion with MRI
5K23NS002147-05Kasner, ScottEarly determination of stroke subtype: ct angiography
5R01NS039325Albers, GregoryNew mri techniques prior to t-pa therapy after stroke
1R01HL082481Culp, WilliamIschemic Stroke Treatment with Microbubbles, Clot Lysis
5R01NS038471Dunn, JeffreyNear infrared/MR system for imaging brain oxygenation
5P41RR015241Hillis, ArgyeReperfusion therapy in stroke candidates using mr perfusion & diffusion imaging
5R01NS047607Moseley, MichaelImproved PWI Methodology in Acute Clinical Stroke
5P41RR009784Moseley, MichaelClinical mr diffusion & perfusion in clincial hyperacute stroke
2R01HL039810Rosen, BrucePerfusion Imaging with Magnetic Resonance
5R01NS038477Sorensen, AlmaMRI Diffusion/Perfusion Mismatch in Human Acute Stroke
5R01NS038477Sorensen, AlmaMri diffusion/perfusion mismatch in human acute stroke
8R01EB002628Thulborn, KeithTissue viability in stroke by sodium mr imaging
2R01EB002634Van Zijl, PeterFunctional Magnetic Resonance Studies of the Brain
1Z01NS002975Warach, StevenSection on Stroke Diagnostics and Therapeutics
5R01NS042607Wityk, RobertInduced Hypertension for Acute Ischemic Stroke
5R29NS036211Wong, EricQuantitative magnetic resonance imaging of perfusion

J. Endpoints: The previous report identified the need for further, rational development of useful stroke trial endpoints as resources.

The 2001 report specifically identified Quality of Life (QOL) instruments as an important needed resource for stroke trials. In the past 5 years, substantial progress in refining and validating stroke-specific QOL scales has occurred, including NINDS-funded studies of the Stroke-Specific Quality of Life Scale and NIA-funded studies of the Stroke Impact Scale. QOL scales incorporate a wider range of patient illness experience in outcome ascertainment and reduce the floor and ceiling effects seen with rater instruments.

Progress has additionally occurred in several other areas of acute stroke trial endpoint analysis. Bayesian techniques have been developed and implemented for identifying the most promising dose in phase 2 trials by utilizing positive (good outcome) and negative (adverse safety outcome) outcomes in real time to select the most informative dose to next be tested, including in the NINDS-supported ROSIE trial. Analyzing the entire range of outcomes ("shift analysis") has been refined and given clinical interpretability. For a disease like stroke that is disabling as well as fatal, incorporating all information from the spectrum of patient outcomes, rather than discarding outcome information by dichotomizing outcome scales, often improves study power. NINDS support has facilitated development of shift analysis (MR RESCUE, FAST-MAG). Baseline severity adjusted endpoint analysis has been refined and implemented in phase 2 and 3 stroke trials, including trials of the a glycoprotein 2b/3a inhibitor, abciximab. By adjusting win criteria to the severity of stroke at entry, this analytic technique increases study power in acute stroke clinical trials.

A variety of useful auxiliary and surrogate biomarkers have been further refined and implemented in phase 2 stroke trials, for more rapid identification of most promising drug dose. Available biomarkers now include the following: 1) For recanalization treatments: transcranial Doppler TIBI scale score, the MR reperfusion ratio, and CTA/MRA noninvasive angiographic assessment of vessel patency (NINDS support of CLOTBUST, DEFUSE); 2) For all acute ischemia treatments: salvage of penumbra defined on diffusion/perfusion MR or CBV/CTP CT imaging (NINDS support of DEFUSE, MR RESCUE); 3) For intracerebral hemorrhage treatment: frequency of hemorrhage expansion on serial CT/MR imaging (NINDS-support of HEME-Surgery, MISTIE, and industry-funded FAST trial).

K. Therapy to improve collateral perfusion: In several places, the original report alluded to devices and strategies that might be used to enhance perfusion into areas of critical blood flow limitation. Many patients with occlusion or critical stenosis of cerebral large arteries do not show recanalization or reperfusion in response to thrombolysis. Some of these patients also fail to show recanalization with embolectomy, angioplasty, or other relatively new interventions aimed to open occluded arteries. In some cases, there are lengthy segments of critical arterial stenosis in the distal internal carotid artery or middle cerebral artery, or both, that cannot be recanalized. Some such patients may show improved perfusion of ischemic tissue by improving collateral circulation.

Temporary induced blood pressure elevation holds promise for improving collateral circulation. This intervention is based on animal studies showing that within the ischemic penumbra, there is a loss of autoregulation, such that increasing systemic blood pressure increases regional cerebral blood flow in ischemic tissue. Recent case series, older series, and a pilot randomized trial have indicated that this intervention may result in improved perfusion and improved function, particularly in patients with large vessel arterial stenosis. An NINDS funded trial (Wytk) is determining whether MR imaging, including DWI and PWI, can identify patients who are most likely to respond to this intervention within 12 hours of stroke onset. In vivo monitoring of norepinephrine-induced blood pressure elevation indicated that increased mean arterial pressure results in an elevation of cerebral perfusion pressure and a selective increase of peak mean flow velocities of the MCA on the affected side

A novel intervention to improve collateral circulation that is currently being studied involves an interventional procedure inflating an intra-aortic balloon to differentially increase blood flow to the ascending aortic artery. The purpose of the Coaxia aortic balloon is to augment cerebral perfusion by occluding 70% of the abdominal aorta below the level of the renal arteries for up to 30 minutes. This reportedly diverts cardiac output to the head and increases cerebral perfusion. This principal has been demonstrated in a pilot randomized prospective study.

L. Medical Devices. Linked to SPRG Recommendations: Randomized trial, MR Rescue, testing efficacy of the Merci Retriever system has been funded by NINDS. The trial also tests the utilization of MRI in selecting patients that may benefit from revascularization therapy in acute stroke.

Not Linked to SPRG Recommendations:

In August 2004, the FDA gave Concentric Medical clearance to market its Merci Retriever system to "remove blood clots from the brain in patients experiencing an ischemic stroke". This approval was granted through the 510(k) process and based on a prospective non-randomized cohort study.{Smith, 2005 #1418} The approval of this device was not linked to SPRG recommendations.

2b. The key unresolved scientific questions identified as such at SPRG in 2001, as well as newly recognized questions.

A. Recanalization. We have not yet found the ideal recanalization therapy. Despite the ongoing trials mentioned above, the need remains to find a way to perform recanalization with minimal risk. Among physicians who are NOT trained in Vascular Neurology, fear of hemorrhage with intravenous fibrinolysis remains the single greatest barrier to implementation of this therapy. Four main avenues may yield success. First, fibrinolytic agents that are safer and more effective are needed. Promising agents include those that are more fibrin specific, such as tenecteplase, desmetoplase, and plasmin-derivatives, ought to produce far less systemic effects, and could cause cerebrovascular thrombolysis with minimal hemorrhage risk. The NINDS funded trial of tenecteplase (Haley) will determine the advantages of this agent. Second, combining fibrinolytic agents with additional drugs active in the coagulation or platelet cascades may improve initial recanalization yield and reduce early reocclusion rates. Further studies of combination recanalization therapies are desirable. Third, combining lytic agents with endothelioprotectants that stabilize the blood brain barrier could reduce the hemorrhage risk. The finding in SAINT 1 that NXY-059 treatment reduced the hemorrhage risk of rt-PA therapy illustrates this approach. Agents active on matrix metalloproteinases show similar promise. Fourth, mechanical recanalization therapies require further development. In the cardiac circulation, primary angioplasty and stenting is superior to pharmacologic recanalization strategies. Primary angioplasty and stenting should be explored for acute cerebral ischemia patients who harbor in situ intracranial atherosclerosis. For patients who have had thromboemboli arrive in the cerebral circulation from a systemic source, clot retrieval, mechanical clot destruction, and other mechanical approaches merit additional development.

B. Neuroprotection. Through continued funding of trials of acute stroke therapy, the NINDS will hopefully help define effective neuroprotective strategies. In particular, combination trials that are difficult and expensive need further NINDS support. Despite the combination trials mentioned above, there remains very little incentive for industry to cooperate in combination trials. Further, the scientific question in such trials is rather trivial, and study sections tend to undervalue pharmacological trials. Nevertheless, the potential public health benefit remains extraordinary, and NINDS should continue to enhance the review process to benefit large combinatorial trials. Statistical methods for improving the efficiency of such investigations exist and should be applied to more stroke clinical trials. Publication bias against negative trials remains an important issue, especially in industry. NINDS in cooperation with an industry roundtable needs to create an environment conducive to the prompt publication of negative trials. Trials designed to exploit and advance the distinctive ability of neuroprotective agents to be given rapidly, before brain imaging, including in the field, should be facilitated. The increasing deployment of recanalization therapy successfully in everyday practice is transforming reperfusion injury from a theoretical to a practical clinical concern. Accordingly, development of neuroprotective agents specifically beneficial against reperfusion injury is now required.

C. Stroke Centers. The effect of primary stroke center designation is undocumented, although the benefits seem self-evident. Nevertheless, a need exists to rigorously determine whether the JCAHO designation enhances the delivery of stroke care. Primary Stroke Center adoption is proceeding steadily, but slowly, with 43 states yet to formally endorse this organization of acute stroke services. There remains a need to develop research infrastructure within many Primary Stroke Centers to facilitate translational research. In SPOTRIAS centers, a requirement that the patient access core treat a minimum of 1 patient per month under 2 hours from stroke symptom onset may also enhance delivery, by incentivizing prompt care. The effect of this SPOTRIAS program requirement remains to be documented. The need for comprehensive stroke centers seems self-evident, but the attendant requirement for prompt transfer of patients from primary stroke centers, or even medic bypass of other medical centers, has social and economic implications in some communities. There remains a need for NINDS to assess the implications of comprehensive stroke center designation. There is not yet a formal certification process for Comprehensive Stroke Centers. Such a process may permit the delivery of advanced endovascular, neurocritical care, and imaging interventions to a greater proportion of the American populace.

D. Pediatric stroke: Data on safety and efficacy of lytics and other therapies in pediatric patients is needed.

E. Emergency Department Investigators and Protocols: First, there is a brief time-window of intervention associated with many therapies, and investigating them can be difficult. Addressing this will require a degree of innovation similar to that in the FAST-MAG trial, in which patients are being treated within 1-2 hours of symptom onset. Second, the time and financial resources required to develop the infrastructure necessary to conduct a multi-center trial for a specific acute intervention for stroke are significant. Experience in the pre-hospital and emergency department setting with acute stroke trial enrollment is not impressive, and the NETT Network is yet to be tested. Most completed trials have had time windows late in the evolution of acute stroke injury. Third, we have not stimulated much interest in the Emergency Medicine community in acute stroke research (with a few notable exceptions). Since the publication of the NINDS rt-PA stroke trial, little progress has been made, and we must determine why acute stroke does not seem to be a priority for the Emergency Medicine community.

F. Telestroke. Although telemedicine is used widely in acute stroke care, it's efficacy in stroke has still yet to be proven in any randomized clinical trial. The ultimate assessment of telemedicine's efficacy in stroke will include an increased number of patients receiving thrombolytic therapy and also an increased proportion of correct decisions to treat or not to treat when utilizing such a system. Telemedicine efficacy in stroke is under investigation in the NINDS funded STRokE DOC telemedicine clinical trial as part of the San Diego SPOTRIAS program.

Additional areas of future assessment and clarification may include establishing technical requirements, assessing cost, facilitating reimbursement, clarifying legal responsibilities, determining telemedicine's role in stroke center designation, assessing the use of telemedicine for clinical trial enrollments and improving stroke patient care throughout multiple locations within the healthcare system.

Expanding upon those technical and clinical requirements detailed above, ISDN bound vs. site- independent systems should also be examined critically. Lack of mobility may cause delay to patient treatment. Site independent systems that minimize delay to system access may be encouraged in the future. Systems currently range in price from $10,000 per bound workstation (not including a $3-12,000 per year cost of networking lines) to $50,000 per site independent system. Hardware and software costs will likely continue to decrease as technologies advance, though costs associated with practitioner consultations and other dedicated resources for management and support still need to be addressed. Reimbursement must be addressed and championed in order for telemedicine to secure a place in the future of stroke care. Similar to the newly established DRG for thrombolysis, care should be taken to ensure reimbursement for teleconsultations. To date, few economic assessments, other than for tele-radiology, have shown success in economic models of telemedicine. Documenting consultations, ensuring appropriate practitioner licensure, and obtaining hospital credentials should help to smooth issues of liability and legal responsibility for telemedicine. This would likely give teleconsultations legal standing similar to standard face to face consultations. A further issue of endorsement may be the Joint Commission on Accreditation of Hospital Organizations (JCAHO), which certifies primary stroke centers. Telemedicine, if effective, could be encouraged as a means to satisfy JCAHO requirements in order to maximize speed and efficiency of stroke patient evaluations in even more facilities. Telemedicine will undoubtedly be used for the purposes of enrolling research participants into clinical trials as well. One hurdle to this use is the general decreased experience or unfamiliarity with performing clinical research by practitioners in remote settings. NINDS should foster research to assess the reliability of telemedicine for clinical trial recruitment and enrollment. Finally, critically assessing the use of telemedicine in various stages and locations of the patient's healthcare (including the ED, inpatient, or even rehabilitation facility) should be addressed and encouraged as well.

G. Reimbursement. The increased reimbursement for acute stroke patients receiving thrombolytic agents, using the new DRG and new ICD-9 Procedure Code, represents a significant advance for stroke centers. Hospitals are now reimbursed for the greater costs associated with caring for these patients and are more likely to provide the resources necessary to establish stroke centers capable of delivering acute stroke care. This will hopefully increase the appropriate treatment of acute stroke patients with thrombolytic agents. However, significant problems remain. The DRG modifications do not address the issue of increased reimbursement for physicians who are required to respond to stroke emergently and spend considerable time with acute evaluation and treatment decisions. Stroke centers also require physician coverage 24/7/365. Additionally, many major stroke centers receive patients in whom IV tPA is initiated elsewhere in a "drip and ship" model. Because the IV tPA is administered at the originating hospital but the patient is admitted to the receiving hospital, DRG 559 cannot be utilized by the receiving hospital because no thrombolytic agent is given there. Thus the stroke center is not reimbursed for the added costs associated with caring for patients receiving the thrombolytic agent. This runs counter to the reasoning for the creation of the new DRG.

H. Diversity. Gender questions have emerged, as pooled TPA datasets suggest gender differences in IV lytic response. Further research into gender differences in response to thrombolytics is required to understand the mechanism of this interaction. Gender differences were also observed in neuroprotection studies, as with Tirilazad showing poorer outcomes in females. There have been too few African-American women in any acute stroke study. In Alzheimer disease, under-representation of minorities in trials was targeted by making funding of ADRCs contingent upon enrolling certain percentages of minorities. Also NCMHD is funding new P20 centers to enhance recruitment of under-represented minorities.

I. Stroke Mechanisms. Research is needed to (1) more rapidly and accurately identify stroke mechanism; (2) identify mechanism-directed therapies, with animal models for different stroke subtypes; and (3) determine whether or not such therapies are differentially effective for separate stroke subtypes. A single multimodal MRI or CT session, imaging the heart, vessels from the heart through intracranial circulation, and brain might prove to be useful in accomplishing the first aim. With regard to the second aim, most current animal models of stroke use the MCA occlusion model; animal models of other stroke subtypes are needed. Advances in serum diagnostics of acute stroke have occurred at only a slow pace and considerable further work is needed to clarify the role of serum biomarkers that may indicate the mechanism of the presenting stroke. Another need exists to determine whether any biomarker, serum or imaging, can differentiate stroke from stroke-mimcs. Considerable effort will be needed to determine whether imaging, with the emergence of multimodal CT and further refinement of multimodal MR, can help to select patients for treatment after the 3-hour.

J. Endpoints. Several new endpoint analytic approaches have advanced over the last 5 years, partly in response to the SPRG. A resulting emerging question is how best to select among competing available approaches to analyzing outcome. Therapeutic success has created new targets for further therapeutic development. Now that reperfusion can increasingly be achieved by pharmacologic or mechanical recanalization, reperfusion injury has become an important clinical entity. Similarly, hemorrhagic transformation, early recurrent ischemia, and late secondary, apoptotic elaboration of injury are important new therapeutic targets requiring endpoint elaboration.

K. Collateral Perfusion. Collateral enhancement is an additional strategy of acute ischemic stroke treatment that has progressed. NIH has funded studies of HHH. Collateral enhancement devices such as Coaxia's Neuroflo have also entered trials. An unresolved question remains, however: Is there really a window of opportunity for late collateral enhancement therapy? Larger, randomized placebo/sham-controlled trials of collateral enhancement interventions are needed to determine efficacy for improving perfusion and function in acute ischemic stroke. Case studies have indicated that patients with large vessel critical stenosis may have extensive brain tissue that is viable but receiving too little blood to function. Such patients seem to respond to such treatments up to a week or more post onset of ischemic symptoms. It is possible that the gradual increase in the degree of stenosis results in ischemic preconditioning that extends the viability of ischemic tissue. Therefore, the time window for effective use of these interventions needs to be determined.

L. Given the lack of appropriate controls in the MERCI trial, there is still considerable debate as to the safety and efficacy of mechanical thrombectomy with the Merci Retriever system for the treatment of acute stroke. The approval of the device, however, has made it extremely difficult to perform efficacy trials where patients may be randomized to a medical treatment only arm.

2c. What needs to be done to address these questions?

NIH needs to clarify current Waiver of Consent issues and achieve harmony among FDA, IRBs, and NIH guidelines. NIH needs to examine the feasibility of centralized national IRB approval for complex multi-center trials. Also, concordance of regulations for indirects and subcontracts is required.

The NINDS needs to continue to actively promote acute stroke care and research in the Emergency Medicine community through education, mentorship and financial support. Measures could be instituted at various points in the time-line (or pipeline) of the development of emergency department investigators. NINDS support of training in both the clinical care of acute stroke as well as research (in the form of training supplements) at a very early time-point in Emergency Medicine clinical training, i.e., residency training, would do a significant amount to increase awareness of the importance of acute stroke research and care in the emergency medicine community. Enhanced NINDS support of training in basic science and clinical research in acute stroke for early-career emergency medicine investigators would significantly aid in the maturation of a cohort of emergency department investigators. An investment in subsequent emergency medicine collaborations with neurological specialists would result in innovations in clinical trial design in acute stroke research, especially for therapies with brief or early time-windows, and allow an investigation of diagnosis and treatment strategies which could be applied very early on in stroke evolution, possibly in the field, by paramedics. It would be advantageous to develop a network of experienced "emergency department stroke centers of excellence" where numerous acute stroke trials could be conducted resulting in expedited patient enrollment and drug/device approval.

--Studies are needed to investigate under which conditions dichotomized analysis, global statistic, responder analysis, shift analysis, and other analytic techniques have differential advantages and disadvantages. Different treatment are likely to have different profiles of impact on outcome, as recanalization agents, cytoprotective agents, and collateral enhancement agents each exert their effects in markedly different ways. Analytic techniques should be selected to optimize detection of clinical relevant, agent-specific treatment effects.

--Imaging biomarkers have been described that correlate with all, and serum biomarkers have been described for several, of the pathophysiologic processes of reperfusion injury, hemorrhagic transformation, early recurrent ischemia, and late secondary, apoptotic elaboration of injury. Refinement of these and development of additional biomarkers of these processes for use in phase 2 trials is critical for the development of the next platform of therapeutics that will minimize acute ischemic brain injury.

Devices are often developed by physicians in collaboration with device companies. Physicians chose to collaborate with companies because they find it difficult to organize large clinical trials alone as they are not sufficiently familiar with the device field, device regulations or the Food and Drug Administration (FDA) regulations. In addition, many physicians often have consulting roles and are on scientific advisory boards of device companies. The collaboration for device development is therefore a natural extension of the already established relationship with the company.

With these kinds of collaborations however, physicians do not have the final decision regarding clinical trial design. Given the rapid and continued changes in technology, it would be impossible to require a clinical trial for each new iteration of a device. Before any given class of devices is adopted for routine use, however, it would seem appropriate that the safety and benefit of the devices for their intended use be proven in a randomized controlled trial. Unfortunately, once the devices are available, it is difficult to complete trials which randomize patients to a medical treatment arm. The difficulty in randomization stems largely from what may be regarded as innocent misperception of the patient or treating physician that use of the device is superior to medical therapy, despite the lack of evidence. Given that the efficacy is a critical concern to the clinical community, new devices should not be used prior to a demonstration of efficacy. These kinds of studies can be encouraged by NIH through seed grants to encourage such collaboration for developing such devices. Furthermore, we encourage companies to perform proper studies proving efficacy, NIH can help support funding to private companies that may not have the capital to perform large randomized trials. Finally, in order to complete trials comparing treatments where there is clinical equipoise, we must remove the barriers to randomization. While provider and patient education is paramount, another strategy is to remove the financial incentive for non-randomization. If CMS and private insurance companies would reimburse only those procedures and devices shown to improve outcome in prospective randomized controlled trials, the interest in completing those trials in a timely fashion would increase. Reimbursement for patients treated within a pivotal clinical trial would further increase the incentive for randomization.

2d. What new key research areas emerged since the first SPRG meeting in 2001?

See above.



Co-Chairs: Michael Chopp, Larry B. Goldstein

Members: S. Thomas Carmichael, Steve Cramer, Bryan Kolb, Randolph Nudo, Jack Parent, Tim Schallert, Samuel Weiss

NINDS Liaison: Eugene Golvanov


1) Basic studies in animal models of experimental stroke have established that focal ischemia induces alterations at the molecular and cellular levels in areas of brain adjacent and distant to the injury. These alterations are associated with neuronal migration to areas of injury as well as axonal and dendritic sprouting, the formation of novel projections, neurogenesis and angiogenesis within both the partially damaged and intact brain and spinal cord.

2) Improved behavioral recovery associated with structural changes in brain has been demonstrated after stroke in a variety of animal model systems (e.g. ischemic and hemorrhagic stroke) with a variety of approaches including cell-based (MSCs, cord blood, stem cells, neurospheres), pharmacological (cGMP, growth and neurotrophic factors, amphetamine, EPO/CEPO, statins), behavioral (enriched environment, voluntary exercise), electrical stimulation, and immune-based interventions.

3) Advances in imaging technology with potential application to the study of brain repair that can be applied in vitro, in tissues and in vivo have include the development of novel MRI sequences and optical imaging techniques (e.g., dual-photon, laser capture microscopy).

4) Recognition of the importance of the neurovascular unit has led to a paradigm shift from a neurocentric to a tissue and cell-cell coupling systems approach to understand brain tissue damage and recovery.

5) Functional sensorimotor and cognitive tests for rats, mice and other animal model systems have been developed to allow for better translational research.

6) Studies show that both aged and young animals can recover after stroke suggesting that brain repair treatments may have similar types of effects in old vs. young brain.

7) Human imaging studies using a variety of techniques (e.g. MRI, PET and TMS) have shown that functional recovery of motoric, sensory, language and attention functions in cortex ipsilateral to a stroke are closely correlated with successful neurological recovery, whereas activation of cortex contralateral to the stroke correlates with poor functional recovery, at least at some time points.

8) DNA-micro array techniques have become available and are being applied to studies of brain plasticity.

Key Unresolved Questions:

1) What are the key cellular and molecular events that mediate behavioral recovery post stroke?

2) Which therapeutic strategies stimulate recovery after stroke?

3) Does the human brain undergo similar restorative changes to that found in experimental animals?

New Key Research Areas That Have Emerged Since Sprg-1:

1) What are the relationships between cellular (e.g., axonal/dendritic sprouting, neurogenesis, angio-genesis, and glial and white matter changes) and molecular events following brain injury and which are causally related to behavioral recovery?

2) What key molecular events regulate post-stroke plasticity and recovery and what are their detailed temporal profiles? What are the interactions between neurogenesis and angiogenesis, and their impact on the brain microenvironment? Which genes and proteins regulate these processes?

3) Which therapeutic strategies most effectively stimulate and enhance recovery after stroke (e.g. cell, drug, environmental enrichment, or combination therapies)? What are the underlying mechanisms that promote recovery of function using cell-based and pharmacological restorative therapies? What are potential adverse effects of these therapies? What is the optimal timing of interventions based on temporal profiles from cell interaction and genomic and proteomic studies? How does the size and location (e.g. cortical, subcortical) of an infarct impact recovery of function and the response to restorative therapies? Can the brain be "primed" to enhance restorative therapy?

4) How does an animal's behavioral experience affect the processes underlying behavioral recovery?

5) Does the human brain undergo the same types of adaptive changes found in animal model systems? How does repair in animal model systems, especially rodents, relate to humans at multiple levels (gene, protein, cell system, and behavior).

6) Can neuroimaging be used to identify patients likely to benefit from specific interventions? How can MRI (e.g., tractography, fractional anisotropy, fMRI) or other techniques be used to identify changes in brain predictive of and related to recovery such as axonal sprouting, angiogenesis, and endogenous neurogenesis in the brains of experimental animals and humans after stroke? Can these techniques be used to monitor the response to therapy?

7) Better characterize the chronic phase of human and experimental stroke, including natural history, imaging, patterns and mechanisms of brain adaptation to stroke and strategies for long-term repair.

8) Support Phase I/II clinical trials to test neurorestorative therapeutic approaches for the treatment of stroke.

9) Evaluate tissue engineering approaches to treatment of post-ischemic brain, including use of scaffolds.

10) Determine genetic factors that affect recovery in stroke patients.

What Needs To Be Done?

1) Standardize experimental and clinical methodologies for neurorestorative research and accelerate their development.

2) Support translational research by fostering interactions between basic scientists, preclinical and clinical investigators involved in post stroke recovery research to better couple preclinical and clinical studies.

3) Support descriptive and collaborative research efforts in post stroke recovery.

4) Support the development of imaging (e.g. MRI) for assessing the processes underlying recovery from stroke .

5) Increase the number of Phase I/II clinical trials for recovery after stroke.

6) Promote the development of consortia for restorative clinical stroke trials.

7) Improve patient enrollment in clinical trials of stroke.

8) Develop a bank of human autopsy material to evaluate neurogenesis, angiogenesis and neuronal reorganization after stroke.



Co-Chairs: Joseph P. Broderick, Robert G. Hart, Karen C. Johnston

Members: Gabrielle DeVeber, Michael D. Hill, Rebecca Ichord, S. Claiborne Johnston, Paul Muizelaar, Arthur Pancioli, Peter A. G. Sandercock, Barbara C. Tilley

NINDS Liaison: Robin Conwit, John Lynch, Claudia S. Moy, Katherine Woodbury-Harris

Clinical trials in stroke are an essential component of the research mission of NINDS. The value of clinical trials in moving experimental therapies into clinical practice is indisputable (Lancet 2006). In 2001, the SPRG identified a number of goals, challenges, and barriers to continued progress in stroke clinical trials. Despite considerable progress since that time, many challenges and barriers remain and are the focus of this report.

Per the format requested by the SPRG leadership at the September 19-20 meeting, this report is organized into four topic areas:

1. Seminal scientific advances in stroke clinical trials

  1. Stroke is treatable beyond the three-hour window (but time is critical).
  2. Neuroprotection remains a viable strategy for acute stroke treatment.
  3. Treatment effects for secondary prevention may vary by ischemic stroke subtypes and baseline patient characteristics
  4. Trials have advanced knowledge regarding the optimal medical and surgical treatment for prevention (including antithrombotic agents, blood pressure control, lipid-lowering, aneurysm coiling etc.).

2. Key unresolved scientific questions

  1. Better understanding of the complexities of timing, dosage, duration, and clinically relevant outcomes for recovery trials.
  2. Prevention of vascular-mediated cognitive decline.
  3. Translation of adult risk/benefit data regarding stroke treatment and prevention to the pediatric population.
  4. Identification of the specific components of acute stroke care that contribute to improved outcomes (from the emergency department to discharge from rehabilitation).

3. Key research areas emerging

  1. Evaluation of the efficacy and safety of devices by randomized trials.
  2. Vascular cognitive decline.
  3. Acute interventions for intracerebral hemorrhage (ICH).
  4. Multi-dimensional standardized outcome assessments.
  5. Multi-modality combination interventions.

4. Top priorities

  1. Support additional clinical trials in:
  2. Address recruitment problems.
  3. Improve trial efficiency:
  4. Additional NINDS Workshops to organize the stroke research community on specific issues and problems. Topics of current importance include:
  5. Encourage and sponsor scientifically valid studies of factors that influence implementation of NINDS clinical trial results into clinical practice

  1. Seminal scientific advances in stroke clinical trials.

Since publication of the 2001 Stroke PRG report, many important stroke clinical trials of acute therapy, prevention, and recovery have been completed and have contributed to clinical and scientific advances in stroke treatment and prevention. In addition, several major new trials have been funded, have begun recruiting patients, and promise additional insights. We identified four areas of particular scientific advances supported by one or more of these trials.

  A. Stroke is treatable beyond the three-hour window (but time is critical).

The pooled analyses of the randomized trials of rt-PA demonstrated that the efficacy of rt-PA is clearly time-dependent and that a benefit for rt-PA likely extends beyond 3 hours (Lancet 2004). The DEDAS and DIAS Phase II trials also indicated that selection of patients by MR imaging may be useful in extending the time window for treatment with IV thrombolytic therapy beyond three hours; a phase III trial based on MR imaging is ongoing. The UK-sponsored IST-3 trial is currently exploring the effect of thrombolysis with rt-PA up to 6 hours after stroke onset.

  B. Neuroprotection remains a viable strategy for acute stroke treatment.

Neuroprotection was demonstrated to be an effective clinical strategy in cerebral ischemia with the publication of two positive treatment trials of hypothermia following cardiac resuscitation.

Subsequently, the phase II SAINT Trial of NXY-059 was the first neuroprotective trial that was positive as determined by one of its primary efficacy endpoints, and an exploratory subanalysis suggested a potential protective effect for prevention of hemorrhagic transformation in patients treated with rt-PA. Unfortunately, preliminary results from the phase III SAINT II Trial indicate that NXY-059 does not improve outcome or protect against hemorrhagic transformation in patients treated with rt-PA..

  C. Treatment effects for secondary prevention may vary by ischemic stroke subtypes and baseline patient characteristics.

It has long been debated whether it is worthwhile characterizing ischemic stroke pathogenetic subtypes when testing treatment interventions. Such characterization is imperfect, expensive, and based on pathogenetic concepts that have not been well validated. Despite these caveats, ischemic stroke subtypes do matter for secondary prevention. The results of the WARSS and WASID trials document no overall benefit of warfarin over aspirin for secondary prevention of stroke in patients with primary cerebrovascular diseases. On the other hand, the NINDS-sponsored Atrial Fibrillation Investigators collaboration has pooled individual patient data from nine randomized trials and has shown very large relative and absolute benefits of warfarin over aspirin in patients with atrial fibrillation who experienced ischemic stroke or TIA. The ongoing NINDS-sponsored Secondary Prevention of Small Subcortical Strokes (SPS3) trial is comparing antiplatelet regimens and even lower target levels of blood pressure control in a tightly defined group of patients with "lacunar" infarcts due to cerebral small artery disease.

In addition, differential efficacy based on patient characteristics, such as sex, age, race, stroke severity, future risk, comorbid disease, etc. may exist and may be a consideration in the design of future clinical trials.

  D. Trials have advanced knowledge regarding the optimal medical and surgical treatment for prevention (including antithrombotic agents, blood pressure control, lipid-lowering, aneurysm coiling etc.).

The NINDS-sponsored WARSS demonstrated the relative equivalency of warfarin and aspirin in secondary stroke prevention in patients with ischemic stroke from a non-cardiac origin. Subgroups of subjects with anti-phospholipid antibodies and patent foramen ovale had similar outcomes with warfarin and aspirin. Because of the greater expense and difficulty of use for warfarin, aspirin is currently seen as the treatment of first choice. The NINDS-sponsored WASID trial demonstrated that aspirin is preferable to warfarin for secondary stroke prevention in subjects with 50% or more atherosclerotic stenosis of an intracranial artery since the prevention of recurrent stroke was similar for both medications and warfarin-treated patients had a higher rate of serious hemorrhagic complications.

The industry-sponsored MATCH and CHARISMA trials demonstrated that low-dose aspirin alone was as effective for secondary stroke prevention as aspirin and clopidogrel combined. Bleeding complications were substantially higher in patients treated with clopidogrel and aspirin as compared to those treated with clopidogrel alone, with no clear increase in efficacy for stroke prevention. The ESPRIT compared extended-release dipyridamole plus low-dose aspirin to low dose aspirin alone, and it reported the combination to be superior. Enrollment in the industry-sponsored PRoFESS study, which compares clopidogrel and extended-release dipyridamole, is completed and follow-up is ongoing. These trials comparing antiplatelet agents for stroke prevention have involved about 40,000 participants at high-risk of stroke and have clarified many issues.

The industry-sponsored PROGRESS trial demonstrated the efficacy of intensive anti-hypertensive therapy for secondary stroke prevention, extending earlier observations from primary stroke prevention. Debate continues whether the benefit is limited to specific classes of anti-hypertensive medications or lowering of blood pressure in general. The ongoing NINDS-sponsored SPS3 trial is testing even further target levels of blood pressure lowering in patients with MRI-defined "lacunar" infarcts, with careful assessment of cognitive function as a key secondary outcome.

Multiple trials of statin agents demonstrated their efficacy in prevention of cardiovascular events, including stroke, in patients with coronary artery disease and coronary risk factors. Recently, the industry-sponsored SPARCL trial confirmed the efficacy of high-dose atorvastatin for secondary stroke prevention. Of note, the risk of intracerebral hemorrhage during follow-up was increased in those assigned high-dose statin; the absolute increase was small, but it established an important proof-of-concept to support findings from epidemiological studies of the relationship of cholesterol and brain hemorrhage.

The U.K. MRC-sponsored Asymptomatic Carotid Stenosis Trial confirmed the findings of the earlier NINDS-sponsored ACAS Trial that supported the modest absolute benefit of carotid endarterectomy in patients with moderate to severe asymptomatic carotid stenosis who were operated upon by surgeons with low operative morbidity.

Several randomized trials have compared carotid stenting to carotid endarterectomy in patients with moderate to severe carotid stenosis with prior ischemic stroke or TIA as well patients with asymptomatic carotid stenosis. The results thus far are mixed with several trials indicating similar morbidity and short-term stroke prevention for carotid stenting and carotid endarterectomy. The ongoing NINDS-sponsored CREST trial is powered to address this comparison definitively.

The International Subarachnoid Hemorrhage Trial (ISAT) demonstrated the superiority of coiling as compared to clipping of intracranial aneurysms in subjects with ruptured aneurysms deemed suitable for either approach, particularly in the posterior circulation. This trial has engendered considerable controversy, and its appropriate clinical application continues to be hotly debated.

Other notable clinical trials that have advanced our concepts of stroke treatment and prevention:

  • An NINDS-sponsored Phase II study of high-dose human albumin demonstrated safety and potential efficacy in patients with acute ischemic stroke, with a suggestion of a stronger benefit in patients receiving rt-PA. The results of this pilot study have led to the ALIAS Phase III trial, which will evaluate the effect of albumin in two parallel cohorts defined by whether or not rt-PA is administered. (Stroke 2006; 37:2107-2114).
  • A Phase II study of magnesium sulfate demonstrated that it is feasible for paramedics to give a potentially neuroprotective drug in the ambulance and that this approach accelerates the start of experimental therapy by 2 hours. The results of this pilot study have led to the FAST-MAG Phase III trial which is evaluating the effect of magnesium sulfate in under 2 hours from onset patients, over half of whom are treated within the first hour of onset.
  • Abciximab (a glycoprotein IIb/IIIa inhibitor) alone was found to be an ineffective treatment for acute ischemic stroke when administered within five hours of onset.
  • The Concentric Retriever device became the first FDA-approved device for clot removal in patients with acute ischemic stroke. Of note and concern to most Committee members, the efficacy was not assessed by a randomized trial, and hence the risk/benefit ratio remains unclear. A randomized trial sponsored by NINDS is underway (MR RESCUE).
  • The paradigm of combined intravenous and intra-arterial recanalization was promising in Phase II pilot trials and is being evaluated in a Phase III Trial as compared to standard IV rt-PA (IMS III).
  • The NINDS-sponsored Vitamin Intervention for Stroke Prevention (e.g. VISP) Trial demonstrated no advantage to high-dose B vitamin replacement in patients with elevated homocysteine.
  • The STOP II trial demonstrated that stopping blood transfusions in patients with sickle cell disease, whose intracranial disease and symptoms had stabilized, resulted in worsening of underlying intracranial vascular disease.

  2. Key unresolved scientific questions.

  A. Better understanding of the complexities of timing, dosage, duration, and clinically relevant outcomes for recovery trials.

Randomized trials of recovery have increased substantially over the past five years although the sample sizes of reported trials remain small. In particular, constraint-induced therapy has been shown to improve the use of the paretic arm as compared to placebo in appropriately selected patients, even years after stroke onset, and the benefit persisted for the two years of the study. Pilot trials of other physical therapy approaches, medications, electrical stimulation of motor cortex, and stem cells are ongoing, but serious, unique methodological issues remain to be resolved. Treatment efficacy may well be time-dependent, with a window of efficacy and, at other times, potentially causing harm.

Further, the need to test complex interventions in rehab trials makes their scientific hypotheses less focused and impairs funding likelihood by review committees. Multi-center (and even multi-national) research networks/groups in rehabilitation and recovery RCTs have not been well-established and would facilitate long-needed trials. The need to develop well validated, widely accepted clinically outcomes is clear.

  B. Prevention of vascular-mediated cognitive decline.

Recent studies reveal that risk factors for developing Alzheimer's disease include the traditional major vascular risk factors. Treatment of hypertension in the elderly significantly reduced dementia attributed to Alzheimer's disease in the SysEur trial, thus supporting a hypothetical interaction of vascular risk factors and small artery vascular disease with degenerative dementias traditionally believed to be non-vascular in pathogenesis. Because vascular risk factors are modifiable with currently available therapies, this overlap warrants additional urgent study.

  C. Translation of adult risk/benefit data regarding stroke treatment and prevention to the pediatric population

When do RCTs that have already been done in adults need to be repeated in children? This is a key unresolved issue that likely must be considered "case-by-case". Because of potential differences in the pathogenesis of vessel occlusion underlying pediatric stroke, the risks and absolute benefits of an intervention in children vs. adults optimally require separate assessment in many circumstances.

  D. Identification of the specific components of acute stroke care that contribute to improved outcomes (from emergency department to discharge from rehabilitation)

Treatment in a specialized acute stroke unit has been demonstrated to improve long-term outcome. However, it is unclear which specific aspects of acute stroke unit care (e.g., early mobilization) contribute to this improvement. The effectiveness of many treatments and procedures to manage increased intracranial pressure, promote resolution of hemorrhage, manage blood pressure and blood glucose, for example, have not been evaluated in definitive trials. This is important to better define the contribution of specific components to optimize use of scarce resources at the community hospital level.

  3. Key research areas emerging.

  A. Validation of the efficacy and safety of devices by randomized trials.

The Committee was concerned about current standards for the evaluation of medical devices to treat and prevent stroke and their subsequent approval by regulatory agencies. Resolving the obstacles to maximizing high quality randomized, blinded, controlled clinical trial data for devices to allow for informed use of such devices must be a priority. There is a perceived "double standard" regarding levels of evidence required for the approval of medical devices vs. of pharmacological treatments. While arguably outside of the scope of the SPRG, it seems relevant to acknowledge that approval by the FDA and CMS prior to the conduct of RCTs makes execution of RCTs more difficult and prolongs the time until RCT evidence that is necessary to determine optimal clinical application is available. In addition, linkage of device testing to reimbursement by CMS should be examined carefully since reimbursement for devices by CMS to hospitals prior to demonstration of clinical efficacy may impede successful completion of key randomized trials. For example, reimbursement could be limited to hospitals that are participating in clinical trials or for humanitarian exemptions in exceptional cases. CMS has done this in the past with the NINDS-sponsored ACAS and NASCET trials.

  B. Vascular cognitive decline.

In addition to clinical and subclinical stroke (i.e., temporally discrete, focal brain lesions), vascular disease of the brain is increasingly appreciated to play a role in progressive cognitive decline in the elderly, albeit through mechanisms incompletely defined at present. Future clinical trials testing interventions aimed at vascular diseases causing stroke should include cognitive assessment and quality of life measures in addition to counting stroke events. Clinically recognized stroke events represent only the tip of the iceberg of vascular injury to the brain.

  C. Acute interventions for intracerebral hemorrhage (ICH).

Treatment of acute ICH has been long neglected, but several important trials have been reported since the 2001 SPRG report. An industry-sponsored phase II dose-escalation trial of recombinant factor VIIa (rFVIIa) for patients with intracerebral hemorrhage demonstrated that rFVIIa may slow progression of bleeding within the first 3 to 4 hours after onset with improvement in clinical outcome. If the efficacy and safety of this treatment is confirmed in the nearly completed phase III trial, rFVIIa would represent the first scientifically proven treatment for ICH. Like, rt-PA, the first FDA-approved therapy for acute ischemic stroke, the efficacy of rFVIIa appears to be time-dependent.

The STICH Trial demonstrated the feasibility of a large multi-national randomized surgical trial of ICH. Evacuation of supratentorial ICH within 96 hours of onset by standard craniotomy was not superior to standard medical therapy. Patients presenting with lobar clots within 1 cm of the surface comprised one exploratory pre-defined sub-group that appeared to benefit from surgical removal. In contrast, those presenting in deep coma (GCS 5 to 8) tended to do better with medical management. Together, the data from both STICH and the other smaller randomized trials suggest that surgery does not appear to be helpful in treating most supratentorial ICH and is probably harmful in those patients presenting in coma. Smaller series of patients treated with minimally invasive approaches have been reported and are promising but have yet to be tested in adequate randomized trials.

An NINDS-sponsored pilot trial is underway to investigate the safety and potential efficacy of blood pressure lowering using nicardipine for acute ICH. A pilot trial found that magnesium sulfate reduced delayed ischemic infarcts evident on CT following subarachnoid hemorrhage. A phase III trial (MASH) is underway and is powered to analyze benefit upon clinical endpoints.

  D. Multi-dimensional standardized outcome assessments.

Outcome assessments for stroke clinical trials now routinely capture such outcomes as quality of life and other patient perspective outcomes in addition to the usual disability and stroke severity outcomes. It has, however, been recognized that there remains a need to include a battery of outcomes for most, if not all, stroke clinical trials that may include disability, functional impairment, neurological dysfunction, cognitive impairment, quality of life, cost and other relevant outcome measures. The development of such a multidimensional standardized outcome assessment for use in the majority of stroke clinical trials has become a priority.

Brief, valid, standardized assessments are needed for incorporation into both prevention and intervention trials. Rehabilitation clinical trials will require appropriate clinical outcomes that are standardized for trial use. The determination of the optimal analysis of outcome measures for a given trial will depend on the individual trial characteristics and distribution of the data

  E. Multi-modality combination interventions.

In the past five years, clinical trials for the treatment, prevention, and rehabilitation of stroke have moved from evaluation of a single or multiple drug treatments to trials that involve combinations of drugs and devices, devices that release drugs, and devices and programs of physical therapy. For example, ongoing acute intervention trials of ischemic stroke include the use of thrombolytic drugs plus clot-removal devices (e.g. Concentric retriever), thrombolytic drugs and ultrasound therapy, and combination of devices. The use of drug-eluting stents, currently used in the coronary circulation, may be considered for occlusive disease in the cerebrovasculature. Finally, several ongoing trials of stroke recovery are evaluating the use of devices, drugs, and cell therapies added to standard or modified physical therapy.

  4. Top Five Priorities.

  A. Support additional clinical trials in:

  •   acute ischemic and hemorrhagic stroke
    Despite scores of recent trials, currently the only approved therapy for acute ischemic stroke is i.v. tPA within three hours of stroke onset. The group recommended strongly that evaluation of treatments for acute ischemic stroke and ICH should continue to be a priority on the basis of huge unmet need and emerging opportunities. The successful SPOTRIAS program should be continued and expanded. If successful in its initial implementation, the NINDS Neurological Emergency Treatment Trial (NETT) Network should be rapidly expanded to include as many centers as possible in order to facilitate the rapid completion of these acute stroke trials.
  •   recovery and rehabilitation
    Clinical trials in this area have proven more challenging than previously anticipated, with variables in dosing, timing and duration adding to their inherent complexity. It is clearly an area of unmet need with large potential benefits. Studies that clarify dosing, timing and duration of recovery and rehabilitation interventions are necessary and will facilitate improved design for recovery and rehabilitation clinical trials.
  •   prevention of vascular-mediated cognitive decline
    As noted above, clinically recognized stroke events represent only the tip of the iceberg of vascular injury to the brain. The NINDS should undertake leadership in organizing large, simple clinical trials for primary prevention and treatment of vascular-mediated cognitive decline (Neuroepidemiology 2005; 25: 91-104). Collaboration with primary care research networks and other NIH Institutes involved in prevention of vascular disease and dementia may be optimal
  •   pediatric stroke

    These areas were selected based on consideration of unmet needs and opportunities.

    Except for the STOP II trial, there have been no completed pediatric trials of stroke prevention, acute treatment, or recovery. The SITT study (Silent Infarct Transfusion Trial) is ongoing in pediatric sickle cell disease. The organization of the International Pediatric Stroke Study, a multi-center consortium interested in pediatric stroke and the publication of large series of pediatric stroke cases, represents an opportunity to determine the feasibility and appropriate methodology for future trials. Needs and opportunities in pediatric stroke RCTs include:

    1. More easily administered and tolerable treatments to prevent stroke in patients with sickle cell disease.
    2. Prevention of stroke associated with cardiac surgery.
    3. Intracranial hemorrhage in neonates.
    4. Secondary stroke prevention in children.
    5. Neuroprotection in childhood stroke
    6. Safety trials of rtPA in children.

  B. Address recruitment problems.

One of the primary impediments to expeditious completion of stroke trials is slow and inadequate recruitment, which results from a variety of obstacles, including lack of organized clinical network infrastructures, lack of qualified and trained clinical investigators, and delays in the process of starting, monitoring, and completion of clinical trials. The barriers also vary with the type of trial (e.g. acute vs. prevention trials). Trials usually underestimate the number of needed sites to recruit the required sample size. Academic centers, by themselves, are inadequate for recruitment of sufficient numbers of subjects for stroke clinical trials. The capacity of academic centers to recruit has been exceeded by the large number of ongoing trials. Proliferation of stroke clinical trials has led to competition for subjects and slows recruitment in all trials. For example, in acute stroke trials, there are currently more ongoing trials within 3 hours of onset than at any previous time.

Successful recruitment depends upon a sufficient number of trained clinical investigators. There are a limited number of stroke-trained neurologists who can serve as clinical investigators and a more limited number of neurologists who are willing to take acute stroke call. In addition, acute stroke patients are first evaluated by emergency physicians who are often less comfortable evaluating and treating stroke patients than they are treating patients with acute cardiac disease. There has been resistance among emergency physicians in treating patients with rt-PA without close interactions with neurologists. Finally, the number of rehabilitation physicians, neurosurgeons, and neuroradiologists with clinical research training is small. This lack of trained clinical investigators and the failure to incorporate clinical research into the clinical stroke care limit recruitment in acute and recovery trials in particular. As a result, neuroclinicians in training do not envision participation in research as an expected part of everyday clinical practice.

Networks of stroke centers and investigators are often trial-based and temporary. Reconstitution of new networks with each new trial is inefficient and slows initiation of new trials. The development of the SPOTRIAS network and the just-starting Neurologic Emergency Treatment Trials Network should better integrate ongoing efforts and may be used as models for other future networks.

Streamlined trial designs could potentially reduce the complexity and cost of multicenter clinical trials. For example, trials could be done in community neurologists' offices. The ongoing NINDS Clinical Research Collaboration is exploring this option. Even at academic centers, research coordinators play critical roles in ensuring the proper conduct of the enrollment, consent, and management of the study subjects. There are also extensive personal contacts needed to facilitate communication with IRBs and institutional officials approving site contracts and consortium agreements that are necessary to start and properly maintain the study. Busy academic physicians are highly dependent upon these coordinators.

In times of decreasing Federal research dollars, the timeliness and success of clinical trials will depend to a large degree on finding novel incentives for investigators to recruit quickly and efficiently. Funding schemes must take into account the pressures on academic faculty at U.S. medical schools. The old academic incentives of the 1980s and early 1990s no longer apply, and NINDS funding must adapt to the new realities.

Finally, distrust by minority populations of participation in clinical research has been an impediment to greater representativeness and diversity of clinical trials.

Table 1. Potential Approaches to Recruitment Delays and Shortfalls

  • Define clearly the problem: NINDS Recruitment Workshop (June 2006)
  • Expand the workforce (community sites)
  • International participation (indemnification issues)
  • Train the workforce (including residents and fellows)
  • Networks with research coordinator support
  • Clinical Research Collaboration (community physicians)
  • Re-examine remuneration strategies (F&A, monetary, academic credit)
  • Collaboration with emergency medicine physicians (NETT Network)
  • Streamlined trial designs
  • Better trial management
  • Central IRBs or IRB collaboration policies
  • NINDS registry of site performance
  • Minority recruitment plans
  • Transparent and ethical processes for assigning stroke patients at sites with competing trials.

  C. Improve trial efficiency

  •   novel biomarkers / surrogate outcomes
    Reliable, validated biomarkers and so-called surrogate outcomes (i.e. non-clinical outcomes) for phase I and II trials, are sorely needed. The potential for decreasing sample sizes and costs make surrogate markers attractive. The potential use of imaging as a surrogate and the use of very early markers for success/failure were discussed as MR imaging has been successful in other diseases (multiple sclerosis, rheumatoid arthritis). Blood-based assays may also have potential. However, the need for surrogate markers to be validated as correlating tightly with and strongly predicting relevant clinical outcomes and responsive to interventions confirmed in positive RCTs was noted. The panel endorsed pursuit of validation of biomarkers as part of phase III RCTs. In the past, poorly planned, expensive studies of potential biomarkers have abused their reputation. Acknowledging the considerable expense sometimes associated with studies of biomarkers due to high-tech aspects, studies of biomarkers might be confined to subsets of participants, and their expense and potential generalizability should be carefully weighed.
  •   novel trial designs and large, streamlined trials
    More rigorous phase I studies and novel and more efficient designs for phase II-III are needed. More frequent adoption of the futility trial approach, as exemplified by the NINDS Parkinson's disease program (NET-PD) and RANTTAS, could be a valuable tool for decreasing the time to identifying treatments not likely to be beneficial and could have a substantial impact on effective use of clinical trial resources. The model of combining phase I and II was discussed. Included in these designs must be consideration of the variability of the population. Better understanding of high risk populations is necessary. The use of risk stratification for clinical trials, factorial design and sequential design were suggested. Risk stratification using individual genetic information must be developed. Developing stroke severity measures that are appropriate for the pediatric population should be a priority as stroke severity is associated with outcome. Consideration of the FDA's perspective was also discussed in regard to new trial design and new outcomes including surrogates. Additionally, large simple streamlined trials may be an option for some clinical trial questions.
  •   Address "process barriers"
    Process barriers include delays in submission, review, and start-up of proposed clinical trials. The interval from submission of an RO1 grant for a clinical trial to randomization of the first patient is typically several years. Regulatory requirements, including NIH and FDA interactions, industry partnerships, IRB, and particularly contracts, all combine to slow study initiation. Establishing subcontracts for clinical sites has emerged as the most common rate-limiting step in study initiation because of the issues of indemnification, intellectual property, institutional control, and economic costs of tests that are outside of the local investigator's control.

  D. Additional NINDS Workshops to organize the stroke research community on specific issues and problems, for example:

NINDS-organized workshops since the 2001 SPRG have been instrumental in advancing clinical trials in several important areas (the Clinical Research Collaboration Workshop, Emergency Neurology Workshop, Recruitment Workshop, Intracerebral Hemorrhage Workshop, Primary Prevention of Stroke Workshop (Neuroepidemiology 2005; 25: 91-104), Neurocognitive Testing Workshop). The Committee believed that resources spent were well invested and applauds the efforts of NINDS staff. Additional workshops are warranted in the areas noted above.

E. Encourage and sponsor scientifically valid studies of factors that influence implementation of NINDS clinical trial results into clinical practice.

Research that focuses on understanding and optimizing the implementation of clinical trial results must be a priority of NINDS to address economic and societal impact issues as well as to standardize clinical trials research methods. Although NINDS has not typically supported research in this area, it is a critical component of the continuum of research in clinical trials. NINDS sponsorship of implementation research for clinical trials will complete the chain from scientific breakthrough to validating new treatments (via clinical trials) to ultimately reducing the disability from neurological disorders. No one is likely to do it as well as NINDS for neurological disorders and NINDS-sponsored research.

Additional Tools And/Or Resources Needed To Facilitate Stroke Clinical Trials.

  • An adaptable, sophisticated NINDS-hosted internet-based trial data management platform , including common data elements, data entry systems with cross-checks, follow-up appointment programs, quality-assurance monitoring was proposed to spare the considerable effort and expense of each new RCT creating its own. A centralized infrastructure associated with a large network of investigators is especially necessary to conduct pediatric clinical trials. In addition, data from existing trials, if could be made available to investigators, could inform decisions regarding trial design and expected outcomes.
  • Training programs for researchers interested in clinical trials would enhance the number of well-trained clinical researchers involved in RCTs. In many other areas of neurology research, there are well-developed training programs beginning at the medical student level aimed at fostering an expert workforce (e.g. medical genetics). This has been lacking in clinical trials. Such a training program is also likely to grow the clinical network and could easily expand to an international training program and network. The current NINDS RFP for such a training program will be critical to achieving this objective.
  • Revised mechanisms of proposing a clinical trial idea to the NINDS may require consideration. Current methods are very time consuming with little feedback early on the Institute's interest in certain areas. In addition to the RFA and RFP mechanisms of soliciting trials based on public health needs, a two-step mechanism of considering investigator initiated ideas as they relate to Institute priorities may be beneficial.
  • Additional research is needed to predict those at high risk for first stroke in order to target populations for efficient trials of primary stroke prevention. About 80% of strokes are first-strokes, and more attention is warranted to primary prevention (NINDS Workshop on Primary Prevention of Stroke, Neuroepidemiology 2005).
  • The importance of assessing the economic and societal impact of NINDS-sponsored RCTs was acknowledged. This is a legitimate consideration when deciding whether to expend public research funds. R01 investigators typically include such projections in their grant applications, but this is done without standard methodology. Several suggestions were put forth by the panel:
    1. Propose a more standard methodology for use by grant applicants, including sensitivity analyses based on a range of possible treatment effects.
    2. Incorporate economic and societal impact formally into the grant review process, including expertise on the study sections.
    3. Require collection of relevant data into the trial protocol to facilitate this to be assessed after completion of the research.
    4. Include prominently such information when RCTs are presented to the NINDS Advisory Council for approval.



Co-Chairs: Marilyn Cipolla, Frank Faraci, Costantino Iadecola

Members: David Busija, Donna Ferriero, Edith Hamel, Donald Heistad, Zvonimir Katusic, Maiken Nedergaard, Mark Nelson, William Pardridge

NINDS Liaison: Tom Jacobs

The following is a summary based on input from our working group. Additional input was provided by solicitation of opinions from other investigators in this field. The presentation is organized primarily based on cell type, focusing on the progress made since the 2001 SPRG and on outstanding issues that remain to be addressed. This analysis is followed by a conceptual summary and final summary points.


Progress . Cerebral endothelium is the main component of the blood-brain barrier (BBB, discussed separately below), but also affects vascular tone, vascular structure, and thrombosis. Endothelial cells have recently been shown to affect neurogenesis and neuronal function. Thus, the impact of endothelial in health and disease is widespread and affects diverse systems that go well beyond the blood vessel.

A priority in the 2001 SPRG, there has been continued progress in understanding the overall impact of cardiovascular risk factors on cerebral endothelium, including the impact of oxidative and antioxidant mechanisms. Reactive oxygen and nitrogen species have emerged both as mediators of endothelial function and dysfunction. Although there has been progress in defining mechanisms of vascular protection and dysfunction at the molecular level, this work is generally in its early stages. Some of these mechanisms may differ both qualitatively and quantitatively from what is described in the peripheral circulation, reflecting the uniqueness of cerebral endothelial cells.

Some barriers described in the SPRG have been overcome. There has been great progress in development of genetically altered mice, including models with endothelial cell-specific alterations. The sophistication of gene arrays has advanced greatly. There has been progress in development of methods to study small intracranial vessels, which was identified as a barrier in the SPRG. Methods for imaging of brain and vessels in mice have advanced. Thus, we now have some of the tools needed for sophisticated studies in mice. Models of stroke-prone hypertensive mice with intracranial hemorrhage have been developed. These models hold clear potential for advancing further our understanding of pathophysiology of hemorrhagic stroke.

Unresolved and New Questions . Development of sophisticated methods to study cerebral vascular biology provides a great opportunity to examine physiological and pathophysiological mechanisms. Some areas that are especially promising for additional study include: mechanisms of endothelial dysfunction, the role of endothelium within the neurovascular unit (NVU), endothelial response to injury, thrombosis and anti-thrombotic mechanisms. Furthermore, attention should be devoted to molecular changes in endothelium that can either protect against or promote ischemic and hemorrhagic stroke, as well as dementias and vascular cognitive impairment. Careful in vivo examination of cerebral endothelial function and dysfunction, coupled with sophisticated molecular biological approaches, will provide new insight into the biology of cerebrovascular disease.

The discovery of circulating bone-marrow-derived endothelial progenitor cells (EPCs) represents a major advancement in understanding of homeostatic mechanisms that may affect cerebrovascular function. The importance of EPCs in repair of injured vascular endothelium and angiogenesis has been demonstrated in peripheral and cerebral circulation of experimental animals. However, the absence of consensus in the literature regarding nomenclature of EPCs requires further studies designed to determine phenotypic characteristics of EPCs including identification of specific and reliable markers. Biology of EPCs as it relates to cerebrovascular disease, including mechanisms of mobilization, homing, and differentiation of stem/progenitor cells in the brain is an area of clear interest.

Pre-clinical studies suggest that stroke patients may benefit from transplantation of EPCs. However, numerous questions remain to be addressed before translation of these findings into the clinical arena. For instance, identification of cell types most suitable for therapeutic application remains to be determined. Additionally, it is unclear whether harvesting of EPCs is obligatory, or whether therapeutic mobilization of endogenous cells may be more effective than delivery of exogenous EPCs. Identification of vascular and neural protective cytokines produced and released by EPCs may help to replace cell therapy approaches with pharmacological interventions designed to mimic beneficial effects of EPCs.

Vascular Smooth Muscle

Progress . Recent studies have further elucidated the roles of key ion channels, calcium signaling, and intracellular signaling mechanisms in cerebral arteries and arterioles. Since 2001, it has become clear that excitation does not simply regulate contractility through changes in membrane potential and intracellular calcium signaling, but that gene transcription is also affected. Furthermore, progress has been made in defining mechanisms of intramural vascular signaling in smooth muscle, but the understanding in brain vessels is limited compared to other vascular beds. It has also become apparent that cerebrovascular smooth muscle, in addition to endothelium and adventitial cells, are important sources of reactive oxygen and reactive nitrogen species, which may play a role both in the normal state and in disease.

Structural changes that can impact cerebral blood flow have been shown to occur in cerebral vessels in multiple animal models of disease (and some in human cerebral vessels) and during aging. Inward vascular remodeling occurs during hypertension and may contribute to reductions in baseline cerebral perfusion and vascular reactivity, potentially contributing to oligemia and dementias. These changes have been linked to oxidative stress, which can produce marked hypertrophy independent of changes in arterial pressure. Complex cerebrovascular alterations occur during aging, in which there is hypertrophy of large and medium size cerebral arteries but atrophy of smaller cerebral arterioles.

Unresolved and New Questions . The mechanisms by which intravascular pressure and agonists depolarize and constrict cerebral arteries remain to be fully elucidated. Although it is clear that ion channels, calcium signaling, and other signaling pathways are altered in disease (and presumably with aging), there are still many fundamental issues to be resolved to develop rationale novel therapeutic approaches in this area.

Our understanding of the mechanisms by which smooth muscle cell excitation regulate gene transcription remain to be elucidated. The field of "excitation-transcription coupling" in cerebrovascular smooth muscle should be an important additional area of investigation to link longer term functional and structural changes to alterations in contractility.

Our understanding of the properties of vascular muscle in intracerebral arterioles is very limited. It is clear that these cells receive signals from astrocytes, neurons, and endothelial cells. The mechanisms by which physical forces affect signaling pathways to regulate arteriolar smooth muscle function are not well defined and may or may not be similar to those operating in larger upstream vessels. The role and composition of ion channels and calcium signaling in these vascular muscle cells are largely unknown. To what extent vascular muscle in parenchyma signals back to endothelium, astrocytes and neurons is not known. Insights into these processes will further our understanding the NVU under normal situations, during aging and disease.

Many disease states affect the morphology of cerebral blood vessels, and mechanisms producing such changes in vascular muscle need study. If structural and functional changes represent early, pre-symptomatic markers of disease, it may be possible to interact directly at this level and obtain beneficial effects.

Neurovascular and gliovascular signaling in health and disease

Progress . There has been significant progress in the understanding of the NVU unit over the last 5 years, particularly in the area of coupling between neural activity and blood flow. These advances have been directly linked to previous SPRG recommendations.

Advances have been made mainly by direct visualization of neurons, astrocytes, and arterioles in brain slices, and, in the case of astrocytes, in vivo using two-photon imaging. The use of detection dyes and caged agents coupled with measurements of arteriolar diameter have provided data suggesting an important role for astrocytic end-feet in modulating microvascular resistance by releasing vasoactive signaling molecules.

Other studies also have emphasized an important role of interneurons in the coupling of neuronal activity and cerebrovascular tone. In addition to an intermediate role for astroglia and interneurons, it is apparent that direct diffusion of signaling molecules such as nitric oxide from neurons is an important component of metabolic/blood flow coupling. The identity of other potentially important vasoactive molecules has also advanced.

Alterations in any of the compartments of the NVU may affect the normally tight metabolic/blood flow coupling of the brain. New data from animal models of Alzheimer's disease (AD) indicate marked changes in cerebral vascular function. Similar to AD, many cardiovascular risk factors affect function of cerebral vessels and potentially neurons and astroglia. Therefore, vascular control mechanisms are potentially disrupted at the neuronal, endothelial, and vascular muscle level by disease. These advances are also directly linked to SPRG recommendations from the 2001 report.

The detrimental role of ß-amyloid on neuronal, glial and vascular function has been well documented, but the exact mechanisms are poorly understood. Vascular oxidative stress induced by ß-amyloid appears as a major culprit in the cerebrovascular dysfunction seen in mouse models of AD. In models of hypertension, diabetes and aging, endothelium- and vascular muscle-specific dilator mechanisms are impaired due to ROS. Thus, ROS and oxidative stress appears to be a final, common pathway for vascular and neurovascular dysfunction, despite differing etiology. The NAD(P)H oxidase system may be a major source of ROS and mediator of cerebral vascular dysfunction in multiple disease models including hypertension, diabetes, and AD.

Endogenous antioxidant mechanisms exert major protective effects on the vessel wall in models of hypertension, vascular inflammation, AD, and aging. Despite these encouraging insights, administration of antioxidants has minimal beneficial effects in clinical trials of vascular and neurodegenerative diseases, suggesting that our understanding of antioxidant processes is still rudimentary and incomplete. Our choice of antioxidant in such trails has certainly been limited and may well have been suboptimal for the problems being addressed.

Unresolved and New Questions . The precise interplay and temporal interactions between afferent neuronal signals and their intermediaries, involving either local neurons or astrocytes, and the exact nature of the vasoactive messengers, need to be better defined. Furthermore, there is insufficient information available to indicate how the NVU is affected by acute and chronic disease states that affect the cerebral vasculature or brain parenchyma.

There is a need to evaluate how findings obtained in tissue slices, translate into the in vivo setting and disease models and determine the respective contribution of different cell types and signaling molecules. Moreover, the apparent discrepancy in findings concerning the ability of astrocytic signals to induce contractile responses needs to be clarified, and the conditions where these two responses can be elicited better characterized. Finally, there is a need to better define signaling mechanisms of propagated or flow-mediated vasodilator responses.

Animal models that would allow the simultaneous visualization of neuronal activation, astrocytic signaling and vascular responses would offer great potential to better identify the cellular and temporal relationships at the NVU. There has been a focus on the use of brain slice preparations to study the NVU. Although the overall goal of this work is important, the approach of measuring vascular responses in tissue slices is problematic with multiple limitations, including the fact vessels and not pressurized or under any tension. Thus, improved imaging methods in vivo, need to be further developed and used.

There is a need to investigate the roles local inflammatory processes and sources of ROS production (NADPH oxidase, cyclooxygenases, eNOS uncoupling, etc) and mechanisms that protect against oxidative stress (superoxide dismutase, glutathione peroxidases, etc) in vascular pathologies. Investigations of mechanisms responsible for vascular degeneration and vascular cell death, as seen in brains of AD patients, need to be performed.

There is a need to increase access to fresh human cerebrovascular tissues from brain banks or other sources, for controls as well as patients with AD, hypertension, diabetes, and other diseases that affect blood vessel functions and structure.

Blood-Brain Barrier

Progress . The model of the BBB has been expanded to more fully incorporate the NVU, with recognition of the complex interactions between multiple cells that form the brain microvasculature. A primary function of the NVU is formation of the BBB, a barrier that impedes influx of most compounds from blood to brain. The BBB is formed by tight junctions between endothelial cells. The functional importance of proteins such as occludens and claudin-5 in promoting integrity of the BBB has been established.

It has also become apparent that the BBB is intact in the first few hours after stroke when neuroprotection is still possible. Therefore, any new therapeutic agent that is proposed to be neuroprotective in stroke must cross the BBB. Based on endogenous BBB transporters, there has been advancement in the development of BBB drug targeting technology to deliver drugs across the BBB. New approaches, including strategies to target endogenous transporters within cerebral endothelial cells have emerged. BBB drug targeting technology continues to improve aiming to deliver small molecule drugs via endogenous BBB carrier-mediated transporters, and large molecule drugs via endogenous BBB receptor-mediated transporters.

The NVU constitutes a major interface between brain and the blood. Recent advances have expanded our view on the function of the NVU beyond the formation of a diffusion barrier. For example, the NVU is a vascular niche that allows neurogenesis in the adult brain. The multiple cell types within the NVU are dynamically responding, and likely capable of modulating events on both sides of the BBB.

The microvasculature is also the site for dense innervation originating from cranial autonomic and sensory ganglia, as well as central sources. Characteristically these nerve terminals contain neuromodulators. The target of neuromodulators may therefore include all cells within the NVU, including endothelium, pericytes, mast cells, and astrocytes. These nerves may play a role in vascular inflammation associated with migraine and may contribute to the vascular alterations associated with AD.

Unresolved and New Questions . Astrocytes cover most of the microvasculature with their endfeet. In vitro studies suggest astrocytes, by release of tropic factors, play an essential role in induction and maintenance of endothelial cell tight junctions. Due to technical limitations associated with studying intact tissue, only culture models have been utilized in the analysis of BBB formation so far. Advances in in vivo imaging approaches combined with gene manipulation will likely provide additional insight into our understanding of BBB induction and maintenance. Astrocytic endfeet also play multiple supportive roles and facilitate the delivery of glucose and possibly lactate to metabolic active neurons.

The discovery of aquaporins in endothelium and astrocytic endfeet increases the complexity of the regulation of water homestasis in brain, especially in diseases that promote edema formation. We do not know what normal hydraulic conductivity is in the brain without aquaporins or under conditions in which they are increased or decreased. The involvement of aquaporins in mediating brain edema is not well understood. In fact, we have little understanding of how the BBB is disrupted to cause vasogenic and cytotoxic edema during stroke. This is a clear area of importance.

Pericytes wrap around the microvascular endothelial cells, providing structural and possibly vasomotor control. Loss of pericytes during development results in microaneurysm formation and spontaneous intracerebral hemorrhages. Pericytes have a contractile function in culture, but the role of pericytes in regulation of the vasculature in vivo is poorly understood. The potential role of pericytes in intracerebral hemorrhage in the adult brain is unclear.

The foundation of brain drug targeting technology is an understanding of the basic science of BBB transport systems. The BBB discovery sciences of the future should be directed at both better understanding of known endogenous BBB transporters and the discovery of new BBB transporters. The latter can be accelerated with the further development of "BBB genomics" and "BBB proteomics," new areas of vascular biology that have emerged since 2001.

Gaining a further understanding of the molecular mechanisms regulating the function of the NVU will be important for deciphering pathological disturbances, including BBB breakdown and edema formation in both cardiovascular and neurological diseases.

Cerebral Vascular Biology In Relation to Stroke During Pregnancy and Early Life

Progress . In relation to stroke and the neonate, there has been progress since 2001. The BBB is not disrupted early in the newborn animal after stroke. There are vascular abnormalities in children with stroke (independent of Moya-Moya or sickle cell). Stroke is common in children with congenital heart disease. There are racial differences in childhood stroke.

In relation to pregnancy, recent progress includes case reports of successful use of tPA for stroke in pregnancy. Diffusion-weighted MRI has shown that eclampsia is associated with edema and such findings raise questions about white matter lesions and eclampsia. TCD technology improvements have demonstrated eclampsia is associated with cerebral hyperperfusion and not vasospasm.

In relation to sex hormones, migraine and stroke, the efficacy of sumatriptans in treating migraine is clear. Although likely not the only mechanism, a working model for migraine involves cortical spreading depression, dilation of dural vessels and the involvement of trigeminal neuropeptides to cause neurovascular inflammation. There has been progress in the identification of genetic mutations that affect the incidence, severity and duration of migraine.

Unresolved and New Questions . There are many questions in this area. How important are genetic factors in predisposing for strokes in newborns and childhood? Recent data that a mutation in a gene encoding a collagen subtype causes hemorrhagic stroke highlights the impact of structural proteins on vascular integrity and gives a striking example of the promise in pursuing this line of reasoning. How much neurogenesis and angiogenesis occurs after stroke in the developing brain? How can these processes be enhanced? What are the best methods to study recovery after stroke in the developing brain? There are no animal models of stroke in childhood so many of the new discoveries, like vasculopathies, have not been studied mechanistically.

What is the role of estrogen in stroke during pregnancy and the increased risk? What is the relationship between migraine and stroke? Does treatment with sumatriptans or other migraine treatments decrease the incidence of stroke in women? Despite the outcome of the Women's Health Initiative, a better understanding of mechanisms of actions of gonadal steroids on cerebral vascular cells, and particularly the endothelium, is needed.

What is the risk of stroke and cerebral venous thrombosis (CVT) during pregnancy and the postpartum state? There is considerable variability in the few studies that have looked at this. A consistent finding is that the postpartum state has increased risk - why? How does normal pregnancy affect the cerebral circulation that may predispose to stroke or other cerebrovascular diseases such as CVT? What is the role of changes in cerebral hemodynamics, genetic reprogramming, structural changes, etc. during pregnancy to stroke and cerebrovascular disease outcome? What is the long-term outcome of women with stroke or other cerebrovascular diseases? What is the relationship between eclampsia and stroke? What causes stroke in eclampsia? Is there a genetic link between preeclampsia/eclampsia, endothelial dysfunction and stroke or other cerebrovascular diseases such as CVT? How does pregnancy and preeclampsia affect the BBB to cause edema and reversible posterior leukoencephalopathy syndrome (RPLS)?

To address such questions, several changes would help. These changes should include improved access to imaging for studies of basic biology, better animal models of preeclampsia and eclampsia as well as childhood stroke and vasculopathy.

New research areas should include: eclampsia as a form of RPLS, the role of aquaporins in the brain and endothelium in mediating cerebral edema formation and seizures, the role of inflammation in stroke in the developing brain, and the role of angiogenesis in recovery and repair. We need better education and information related to eclampsia, i.e., eclampsia is not vasospasm.

Overall Concepts

Within the neuroscience and stroke community, there has been an enormous but appropriate effort toward understanding neuroprotection in stroke and producing neuroprotection with therapeutic approaches. A similar effort is needed for the study of vascular protection and its translational application. Such insight may hasten approaches that could prevent, delay, or stop the progression of cerebral vascular disease that contribute to many pathologies including strokes, dementia, and vascular cognitive impairment. The potential impact of such a strategy could be very large as it should be easier to prevent stroke and lessen the development of cognitive decline than to reverse these events after they have occurred.

Numerous studies now suggest that common mechanisms may underlie cerebral vascular abnormalities in multiple diseases. The greatest impact of future research including translational efforts may arise from a focus on such mechanisms. For example, oxidative stress and inflammation appear to play a major role in many disease states. A better understanding of molecular mechanisms that promote and protect against vascular oxidative stress is needed. Recent data suggest that endogenous vascular protective mechanisms have a major impact within the vessel wall. Better understanding of these mechanisms may help in the design of more effective therapeutic approaches. These efforts should include study of transcription factors or pathways that have the potential to affect entire cluster of genes and thus mediate vascular protection at multiple levels. The renin-angiotensin system remains an attractive area for study with great potential benefit. In addition to its well recognized role in hypertension, this system may contribute to vascular abnormalities in atherosclerosis, diabetes and insulin resistance, hyperhomocysteinemia, and aging (and perhaps others).

To date, most studies in this area have used models of single cardiovascular risk factors in young (adult) animals - often only male animals. To better mimic what often occurs in humans, insight from models of combined risk factors and both genders is needed. For example, although hypertension occurs most often in older individuals, the impact of hypertension is normally studied in young animals. Effects of hypertension in models of aging and the mechanisms involved have not to be pursued to any significant degree. Despite the fact that aging has an enormous negative impact on the cerebral circulation, we know very little about the underlying changes in vascular biology with aging per se. Similarly, changes in cerebrovascular biology in the neonate and during pregnancy as it relates to stroke and cerebrovascular disease is largely unknown as well.

All segments of the vasculature continue to need study. For some disease states, an emphasis on the study of cerebral microvessels is appropriate. Other diseases may primarily impact larger cerebral arteries or the entire vasculature. In relation to regulation of cerebral vascular resistance, one of the two major determinants of blood flow, small vessels within the brain parenchyma (including those contributing to the NVU) comprise only about one-third of the brains total vascular resistance. Although the NVU is the major site for transport across the BBB, the BBB is also present in vessels upstream and downstream from the NVU. Venules, for example, appear a major site of disruption of the BBB during hypertension. With the unique distribution of vascular resistance in brain, the coordination of vasomotor responses between different vascular segments (mechanisms such as propagated or flow-mediated vasodilation) is critically important in brain. However, our understanding of such mechanisms and how they change in disease and with aging is rudimentary.

Structural changes that can impact cerebral blood flow are known to occur in cerebral vessels in multiple diseases and during aging. Studies of the impact of cardiovascular risk factors and aging on vascular hypertrophy, atrophy, and remodeling (and the molecular mechanisms involved) should be a focus. For example, although inward vascular remodeling occurs during hypertension and may potentially contribute to oligemia and dementia as well as larger infarcts following ischemia, we known almost nothing regarding the mechanisms involved.

Although genetically altered mice are used widely for studies of stroke and have provided great insight recently into peripheral vascular biology, these models are not used widely for studies of cerebral vascular biology. The full potential of the use of such models for detailed mechanistic studies has not been reached. There is a need to expand their use but also to incorporate the most modern models as they become available. For example, combining human genetics with detailed physiological and pharmacological studies of mice that express human genes and human gene mutations represents a powerful experimental tool that is barely utilized at present.

Other molecular approaches need to be used more commonly in this field. There has been great progress in relation to understanding and using RNA interferance as an experimental tool. Although the approach is used widely elsewhere, it use has been extremely limited for studies of cerebral circulation. Although microarrays continue to improve and their use in now relatively common, they have been used to a very minor extent in cerebral vessels - perhaps in part due to the smaller amounts of tissue available. Despite these limitations, well designed and appropriate use of microarrays could help define clusters of genes and pathways that impact vascular disease in either a positive or negative manner.

The link between cardiovascular risk factors and cognitive dysfunction is very strong and it seems very likely that vascular disease contributes to some forms of cognitive dysfunction. Studies designed to determine whether experimental manipulations that produce vascular protection also affect cognition, and to determine if protection from cerebral vascular dysfunction (or changes in vascular structure) correlates with protection from cognitive impairment may be very worthwhile.

Much of the most detailed and novel work in vascular biology is performed in laboratories that do not study cerebral vessels and typically use models such as rat or mouse aorta or human vein endothelial cells in culture. There is a need for training of investigators who will first learn, and then utilize the most modern and novel approaches for research on cerebral circulation. In particular, this includes a combination of in vitro and genetic approaches with in vivo models that closely mimic the human condition. There is also a continuing need in relation to training of pathologists who focus on neurovascular pathology and cerebral vascular biology.

Summary Points

Key Scientific Advances Since SPRG 2001

  1. Cellular bases of vascular regulation and neurovascular signaling is better defined, in part because of new approaches
  2. Stroke risk factors disrupt cerebrovascular function
  3. Cerebrovascular dysfunction is associated with cognitive impairment
  4. Common mechanisms underlie vascular abnormalities of diverse etiologies
Unresolved Questions
  1. Impact of disease on the neurovascular unit including link to cognition and translational approaches
  2. Protective and destructive roles of endothelium in vascular disease
  3. Mechanisms that mediate changes in cerebrovascular growth and structure
  4. Neurogenesis and angiogenesis during development and after stroke
What Needs to Be Done?
  1. Use or develop animal models expressing human genes and gene mutations
  2. Use or develop models with cell specific gene manipulations, caged compounds, etc.
  3. New imaging approaches in vivo and ex vivo
  4. Better access to human cerebral vascular tissue
  5. Training - apply state-of-the art approaches for study of vascular biology to cerebral circulation
Key Research Areas
  1. Cerebrovascular dysfunction and cognitive impairment
  2. Impact and mechanisms of combined risk factors on cerebral blood vessels, interaction with gender, development and aging
  3. Mechanisms of vascular protection: therapeutic potential
  4. Molecular basis of brain water homeostasis and edema
  5. Drug delivery through the BBB



Co-Chairs: Ralph Sacco, Brett Kissela, Lynda Lisabeth

Members: Robert Brown, Cheryl Bushnell, John Cole, Heather Fullerton, Philip Gorelick, George Howard, Edgar Kenton, Peter Rothwell, Philip Wolf

NINDS Liaison: Richard Benson, Claudia Moy, Katherine Woodbury-Harris

There is a continuing need for epidemiologic research into the distribution, causes and sequelae of stroke. Well designed studies in defined populations have led to a better understanding of the burden of disease, potential targets for intervention, and improvements in clinical management and prevention. However, many questions remain; these priorities are the focus of this report.

The report is organized into three broad areas:

  1. Scientific advances in stroke epidemiology
    1. New evidence concerning the public health burden of stroke across subpopulations of special emphasis.
    2. Translation and validation of epidemiologic findings regarding risk factors through clinical trials.
    3. Evolution of knowledge about traditional risk factors and identification of novel risk factors including data on subclinical disease.
  2. Key unresolved questions
    1. What is the stroke risk in specific subpopulations, including children, and what are the effects of traditional and novel risk factors?
    2. What are the frequency and determinants of stroke outcomes, including recovery and cognitive impairment?
    3. How can we improve the classification of and refine stroke subtype phenotypes?
  3. Top Priorities in Stroke Epidemiology
    1. Integration of epidemiology studies with common data elements, definitions, sharing plans, and central multi-dimensional repositories including imaging, blood and tissue.
    2. Extension of epidemiology studies to provide more robust longitudinal, secular, and temporal trend data.
    3. Increase the opportunities for training in stroke epidemiology and clinical research methods and encourage multidisciplinary translational research collaborations to include epidemiologic expertise.

  1. Scientific Advances in Stroke Epidemiology

A great deal of progress has been made in stroke epidemiology during the interval since the last Stroke Progress Review Group report was issued in 2002. We will try to place the scientific accomplishments in framework of the Scientific Priorities from the last report.

  A. New evidence concerning the public health burden of stroke across subpopulations of special emphasis.

Significantly more knowledge exists today about the public health burden of stroke, especially within subpopulations such as race/ethnic minority groups, women, and children. The advances detailed below are accomplishments relevant to the previous Priority #1 ("Characterize the public health burden of stroke and establish subpopulations for special emphasis").

  • More extensive data are available to characterize the burden of stroke in African-Americans. Risk for incident stroke has been shown to be substantially increased when compared to Caucasians, especially in younger age groups. Case-fatality rates do not differ between these two race/ethnic groups.
  • Data has emerged about the burden of stroke in Mexican-Americans, the largest subgroup of Hispanic Americans, showing that this minority population also has higher risk for incident stroke as compared to Caucasians, especially at younger ages. Mexican-Americans have lower case-fatality rates after stroke, but have a higher risk of stroke recurrence.
  • Data about temporal trends in stroke show that incidence remains unchanged over the last decade despite improved treatment options for risk factor reduction. Stroke mortality has been stable or is declining over this same time period.
  • Women make up the majority of deaths from stroke given their increased lifespan. Risk factors differ between men and women; women are more likely to have atrial fibrillation and hypertension, whereas men are more likely to have coronary artery disease, peripheral arterial disease, and a history of tobacco smoking. Women appear more likely to benefit from some treatments such as intravenous tPA for acute ischemic stroke and aspirin for stroke prevention. Women are more likely to have poor functional outcomes after stroke than men, which may in part be due to older age at onset. However, several studies with age-adjusted outcomes have demonstrated worse outcomes in women.
  • More data have accumulated demonstrating that persons with lowest socioeconomic status are more likely to have a stroke than those of greater social support and financial means.
  • Data about stroke in children has increased substantially since the initial report. African-American children have increased risk for ischemic and hemorrhagic stroke even after controlling for sickle cell disease and trauma. Among otherwise healthy children, with no prior diagnosis of sickle cell disease or congenital heart disease, approximately 80% have an arterial abnormality (such as large artery stenosis) on vascular imaging such as magnetic resonance angiograms. The etiology, or etiologies, of these vascular abnormalities remain unknown; it is unclear whether these lesions are congenital or acquired. As many as 25-33% of children with arterial ischemic stroke will have a recurrent event in 3-5 years. Predictors of recurrence include lack of antithrombotic treatment, coagulation abnormalities, and arteriopathies or other abnormalities on vascular imaging. Perinatal stroke is more common than previously thought, and is associated with maternal risk factors, particularly ones related to inflammation such as preeclamsia, prolonged rupture of membranes, and chorioamnionitis. While TCD screening is part of a proven strategy for primary stroke prevention in children with sickle cell disease, a recent survey demonstrated multiple obstacles to obtaining TCDs in these children, including poor patient adherence and distance to vascular laboratories.

A national/population-based surveillance system was recommended in the previous report, which has not occurred. While a nationwide surveillance system has some advantages, the high cost and the potential limitations of such administrative data sets in terms of quality and quantity of data collection are disadvantages. Some questions can be initially addressed by administrative datasets, but these are often too limited in scope to fully assess incidence of stroke. A variety of population-based, cohort, case-control and registries have been developed to help address this priority. These stroke epidemiologic studies include, but are not limited to: Framingham, Atherosclerosis Risk in Communities (ARIC), Cardiovascular Health Study (CHS), Multi-Ethnic Study of Atherosclerosis (MESA), Northern Manhattan Study (NOMAS), Greater Cincinnati Northern Kentucky study (GCNK), Rochester, MN, Reasons for Geographic and Racial Differences in Stroke (REGARDS), Brain Attack Surveillance in Corpus Christi (BASIC), Baltimore-Washington Stroke Study, New York Islands AVM Study, Alaskan Native Stroke Registry, International Study of Unruptured Intracranial Aneurysms, and the Familial Intracranial Aneurysm study. Moreover, several large registry studies/programs have been started in the interval, including the Paul Coverdell Stroke Registry and the "Get With the Guidelines" program. In addition, the International Pediatric Stroke Study, a large-scale international childhood stroke registry, was initiated in 2002.

  B. Translation and Validation of Epidemiologic Findings Regarding Risk Factors Through Clinical Trials.

Epidemiology has been important in determining the general health impact of certain factors that are commonly translated into clinical care without clinical trials (such as recommendations to cease smoking or increase physical activity after stroke as forms of stroke prevention). Furthermore, epidemiology is of crucial importance in determining points of intervention for future clinical trials, or appropriately focusing interventional educational projects. The advances detailed below are accomplishments relevant to the previous Priority #3 ("Integrate epidemiology into clinical management and prevention.")

  • Public health messages about stroke are now specifically targeted towards the groups at highest risk, including race/ethnic minorities. These messages are culturally sensitive and are thus of higher impact than more generic messages.
  • Public health campaigns related to specific risk factors have been successfully implemented, such as the "Make the Link" campaign to help those with diabetes understand their risk for macrovascular events such as stroke (sponsored by the American Diabetes Association and the American College of Cardiology).
  • Clinical trials have been successfully implemented looking at treatment of hypertension and hypercholesterolemia (established risk factors for stroke), and showing that stroke risk can be reduced by even small reductions in blood pressure and by medications such as HMGCoA Reductase Inhibitors.
  • Epidemiologic data continues to emerge with regard to behaviors in the population that affect stroke incidence and prevalence, and changes over time continue to be documented. For example, high cholesterol and diabetes are increasing in prevalence, and medical treatment for these conditions is increasing. Knowledge of population behavior relevant to stroke is necessary for appropriate allocation of public health spending.
  • Finally, epidemiologic data about management/patient care has been updated since the last SPRG (including use of tPA, preventative medications, utilization of procedures, utilization of neuroimaging, and utilization of rehabilitation). Dissemination of medical practice trends over time must continue to appropriately allocate medical spending, directing interventions where necessary, as well as targeting public health education.

  C. Evolution of Knowledge About Traditional Risk Factors and Identification of Novel Risk Factors Including Data on Subclinical Disease.

Novel risk factors continue to be explored, as well as the specific effect of these risk factors on subpopulations. The advances detailed below are accomplishments relevant to the previous Priority #2 ("Establish the new determinants of stroke and its consequences and identify subgroups with varying risk.")

  • The Women's Health Initiative reported an increased risk of ischemic stroke with combination conjugated equine estrogen/progestin or estrogen alone. The reasons for this finding in apparently low-risk women are unclear.
  • Diabetes, hypertension, atrial fibrillation, cardiac disease, physical activity and other traditional risk factors have been further studied, and their differential impact upon various subpopulations reported.
    • A greater attributable stroke risk for hypertension and diabetes has been found for black and Hispanics compared to whites and the greater attributable risk of atrial fibrillation and cardiac disease for whites.
    • Ischemic stroke patients with diabetes are younger, more likely to be African American, and more likely to have hypertension, MI, and high cholesterol, than nondiabetic patients. Age-specific incidence rates and rate ratios show that diabetes increases ischemic stroke incidence at all ages, but this risk is most prominent before age 55 in African Americans and before age 65 in whites. It has been estimated that 37-42% of all ischemic strokes in both African Americans and whites are attributable to the effects of diabetes alone or in combination with hypertension.
    • Finally, the metabolic syndrome (a co-existence of multiple risk factors) has been shown to be increasing in prevalence and is associated with higher risk for stroke.
  • Novel stroke risk factors have been associated with stroke, such as homocysteine, obesity, left atrial enlargement, left ventricular mass and geometry, brachial endothelial reactivity, irregular carotid plaque, and inflammation and inflammatory markers. Furthermore, subclinical diseases such as carotid plaque thickness, intimal media thickness, white matter hyperintensities, and silent infarcts have been shown to be vascular risk markers.
  • Familial occurrence of stroke, ischemic as well as hemorrhage has recently been more firmly established. Several candidate genes have been evaluated for association with stroke including: 1) phosphodiesterase 4D, the recently isolated STRK1 locus gene, 2.) arachidonate 5-lipoxygenase activating protein gene 3.) thrombomodulin, a key protein in the coagulation pathway; and 4) endothelial nitric oxide, a potent vasodilator. Specific variants within each of these genes have been associated with ischemic stroke.


Key areas of research in stroke epidemiology that should be addressed over the next several years include the following:

  • an understanding of the stroke burden in the US including measures of incidence, case-fatality and mortality overall, by stroke subtypes, temporal trends, and within subgroups of the population,
  • an understanding for the reasons for disparities in stroke risk and outcome among minority populations including but not limited to risk factor profiles, genetic determinants, gene-environment interactions, socioeconomic factors and social determinants of health,
  • a better understanding of stroke in women including sex differences in perception of stroke risk, diagnosis, treatment, and prevention with a goal towards targeted educational messages, the role of inflammatory markers and stroke risk, mode and duration of hormone therapy (HRT) and stroke risk, and preeclampsia and stroke in pregnancy, and
  • the identification of factors influencing stroke outcome, including mortality, severity, recurrence, functional status, quality of life, and cognitive and behavioral endpoints with an emphasis on research that can better inform the development of, and evaluation of, interventions to maximize outcome among stroke survivors,
  • additional national and international pediatric and neonatal stroke studies including studies of childhood cerebral vasculopathies and inflammation among others.

Key broad research questions for future study are outlined below as well the scientific barriers to accomplishing the work.

  A. What is the stroke risk in specific subpopulations, including children, and what are the effects of traditional and novel risk factors?

Data on basic stroke epidemiology are still lacking in some race-ethnic subgroups, such as American Indians, Alaskan natives and Asian Americans, with only limited data available from vital statistics. There is a need to elucidate which race-ethnic minority groups require special research emphasis due to a poor understanding of stroke in these populations. Such research will help to identify any large but currently unknown issues in stroke in these subgroups of the populations.

There is a lack of uniformity across research studies in defining race-ethnic subgroups, which makes comparisons across studies and the combining of data from different studies difficult. Common definitions of race and ethnicity include those used in the US Census mandated by the government. These classifications of race-ethnicity may be inadequate and do not take into account the heterogeneity within race-ethnic subgroups, cultural distinctions, and/or heritage. There is a need for the development of consistent and meaningful definitions of race-ethnic categories to facilitate epidemiologic research within and across subgroups of the population, which is an emphasis of the research priorities in stroke epidemiology. As research in this area moves forward, it will be important to consider the changing US demographics and how this may impact the need for the classification of new race-ethnic groups.

Investigations on the determinants of stroke, particularly in specific populations are needed. Issues that need to be addressed include: a.) establishment and description of new and novel determinants of disease; b.) identification of specific subgroups at differential risk; c.) better understanding of existing ('traditional') factors; and d.) improved understanding of the role of genetics and gene-environment interactions in stroke"). Such stroke determinants include traditional vascular risk factors, behavioral factors, socioeconomic factors, stroke triggers, risk factor clusters, risk factor scores that evaluate multiple domains, subclinical diseases, biomarkers, infection, inflammation, coagulation, family history, dietary factors, and genetic determinants. Comprehensive evaluations are encouraged to include measures of association (relative risk, odds ratio, hazard ratios), prevalence, population attributable risk, and secular and temporal trends. Special emphasis should be placed on factors that are modifiable or treatable to encourage the natural progression from risk factor identification in epidemiologic studies to design and conduct of clinical trials and intervention studies that evaluate innovative approaches to reduce stroke risk.

  B. What are the frequency and determinants of stroke outcomes, including recovery and cognitive impairment?

Epidemiologic studies are needed to evaluate the frequency and determinants of stroke outcomes including recurrence, severity, mortality, vascular cognitive decline/impairment, dementia, functional outcomes, recovery, post-stroke depression, and overall quality of life. Furthermore, it is crucial to utilize clinical epidemiology to characterize the biological process of post-stroke recovery and to determine factors or practices that impact recovery. In this manner, epidemiology can inform future basic/translational research and clinical trials that may improve the outcome for patients after stroke. Efforts are encouraged to characterize the consequences of stroke across the age span including among children, women, and high-risk racial and ethnic subgroups. Studies are encouraged to address issues of economic impact (direct and indirect costs), care-giver burden, health care services, stroke impact on family quality of life, and institutional utilization such as rehabilitation and nursing homes. Specific outcomes need to be tailored to the study population. For example, evaluation of school performance will need to be addressed in studies evaluating the impact of pediatric stroke. Attempts to track measures of severity and relate them to temporal changes in risk factors, medication compliance, and adherence to behavioral modification are encouraged. Studies to identify predictors of recovery are encouraged to establish the preliminary data prior to designing clinical trials or other translational studies.

Pediatric populations pose special challenges. Because of the relative lower incidence of childhood stroke, international studies are needed to gain a better understanding of childhood stroke etiologies, acute stroke complications, neurologic recovery, recurrence rates, and risk factors for recurrence.

  C. How can we improve the classification and refine stroke subtype phenotypes?

Stroke is heterogeneous, comprising a number of disease subtypes which may differ with respect to risk factors, prognosis, outcome and management. Distinguishing among stroke subtypes remains problematic with limitations to currently used classification schemes. There is still a need for accurate stroke subtype-specific epidemiologic data on risk factors, outcomes and secular trends in disease burden. To accomplish this research, there is a need for better stroke subtype classification systems that work across various research studies and that would allow for meaningful inferences within subtypes. Such a classification system would promote research on how risk factors differ by subtype, secular trends in stroke incidence by sub-type, and the identification of heritable cerebrovascular phenotypes to inform more powerful genetic studies as a few examples.


  A. Integration of epidemiology studies with common data elements, definitions, sharing plans, and central multi-dimensional repositories including imaging, blood and tissue

A variety of population-based, cohort, case-control and registry studies have been developed to characterize the public health burden of stroke in the US. The combining of data from these multiple data sources and studies would allow for more powerful analyses to address some of the key unresolved questions in stroke epidemiology. To facilitate this work, there is a need for common data definitions/elements, including definitions of covariates, risk factors and outcomes, across epidemiologic and clinical stroke studies. Support for data sharing across studies should be encouraged through centralized data repositories, multiple PIs on grants, and other means. Efficient data collection through linkage of study data with electronic medical records and administrative databases should also be explored as the country moves towards this technology.

Currently, there are limited NINDS or NIH sponsored tissue banking facilities for stroke specimens. To facilitate the translation of results from basic science into clinical and human reality, NINDS or NIH sponsored tissue banking facilities for human stroke specimens are needed with a focus on fresh, living tissue. Tissue banking would facilitate proteomic, SNP, array and biochemical analysis of tissue as well as in vitro studies in the future. This effort would require the collaboration of investigators across various research studies.

  B. Extension of epidemiology studies to provide more robust longitudinal, secular, and temporal trend data.

The 2002 SPRG report recommended the development of surveillance systems and other research efforts to establish incidence rates for stroke. While such efforts are ongoing, data collection efforts that include the ability to evaluate secular and temporal trends should also be expanded. Issues which will need to be addressed include definitions of common data elements, central data repositories, innovative statistical ways to manage and share data in web-accessible formats, easier access to merging data sets to address questions of interest to different investigator groups, and acceptable models of publication access policies.

  C. Increase the opportunities for training in stroke epidemiology and clinical research methods and encourage multidisciplinary translational research collaborations to include epidemiologic expertise.

To address many of the Stroke PRG research priorities, an enhanced work-force is needed. More opportunities to train individuals in epidemiologic and clinical research methods are needed through the NIH-funded stroke fellowship/training positions (T32, F- and K- awards, etc.). Enhancement of research curricula in neurology and other residency training programs are also needed to expose individuals at the early stages of career development. Efforts to improve the research infrastructure and reduce potential barriers to clinical research are greatly needed which include standardized approaches to educating institutions and IRBs about issues related to HIPAA, genetic studies, sharing of data on research subjects, and the complicated relationship of medical billing and research protocols. Novel approaches are required to interact with practicing neurologists and other healthcare workers in the integration of practice and epidemiologic and clinical research. Community engagement and education efforts are also necessary for population-based research and improved recruitment into clinical research studies particularly among high-risk, under-represented minority groups. Approaches to encourage multi-disciplinary collaborations across epidemiologic, biostatistical, genetic and other disciplines will lead to enhanced research teams with an increased chance of successfully addressing new research priorities. Such collaborations are needed in the design of clinical trials and also lead to better translational research opportunities between basic and clinical scientists. There is a need to train and encourage epidemiologists and biostatisticians into the area of stroke research.



Co-Chairs: Karen Furie, James Meschia, Andrew Singlton

Members: Joseph Broderick, Myriam Fornage, Steve Greenberg, Terry Manolio, Steve Pavlakis, Jonathan Rosand, Ralph Sacco, Dan Woo

NINDS Liaison: Katrina Gwinn-Hardy, Gabrielle Lablanc, John Lynch

Scientific Advances

Since the 2001 Stroke Progress Review Group (SPRG), the field of stroke genetics has advanced in several tangible respects. There is now compelling evidence supporting the heritability of several subclinical stroke phenotypes. Discovering the genetic determinants of these subclinical phenotypes represents an achievable goal because these phenotypes can be analyzed as quantitative traits. Furthermore, they can be assessed with non-invasive techniques that are widely available in the clinical setting such as carotid ultrasonography and magnetic resonance imaging.

There are also emerging breakthroughs in the pharmacogenomics of stroke prevention. Two critical genetic determinants of warfarin dose, VKORC1 and CYP2C9, have been discovered and subsequently confirmed in multiple studies. Variations in these genes account for about 50% of the inter-individual differences in warfarin dosing with other important factors including age, sex and concomitant medications. These discoveries make the prospect for individualized dose-initiation of warfarin a tangible reality for patients within the next 5 years. Such individualized dosing offers the prospect of substantially widening the therapeutic index of one of the most effective stroke prevention therapies presently available.

The favorable performance of studies of other complex genetic disorders represents an indirect advance for stroke genetics. Recent studies have confirmed that common genetic variants contribute to risk of common diseases, including type 1 diabetes, type 2 diabetes, age-related macular degeneration, rheumatoid arthritis, and systemic lupus erythematosus. These discoveries in other complex diseases, in tandem with the demonstration by the Icelandic DECODE study that large-scale genetic studies of stroke are feasible, show the way forward in terms of study designs that will be crucial for success in finding stroke genes.

A compelling example of the potential of genetic information in predicting stroke outcomes was demonstrated in a recent study of children with sickle cell anemia. SNPs in 12 modulator genes in sickle cell anemia were shown to predict the risk of stroke with 98.2% accuracy. This type of data may provide insights into novel approaches for prevention. Impressive as these findings are, they are best viewed as proof of principle requiring independent replication. The ability to use simple pre-symptomatic genetic testing to predict stroke in sickle cell anemia with such remarkable accuracy is not merely academically interesting, it is of vital importance to public health. Gene testing has the potential to supplant current imaging and clinical approaches to risk stratification and open vistas of ultra-early primary stroke prevention.

Tools for empirical gene discovery are becoming ever more accessible, thanks largely to the Human Genome and HapMap Projects. The public release of HapMap data has dramatically improved the efficiency of genetic testing, reducing the number of SNPs required to examine common sequence variants across the entire genome from 10 million SNPs to less than 1 million SNPs. This advance will make whole genome association studies for specific stroke phenotypes feasible. Furthermore, it will allow an unbiased search for the many disease gene variants that exist beyond coding regions and other pre-specified "candidate" regions. The expansion of computerized tools and analytical approaches to complex genetic analyses has improved significantly in the last five years.

There have been numerous accomplishments in the last five years in the field of stroke genetics as a direct result of research funded by the NINDS. Although by no means exhaustive, the following are some of the highlights:

  • A region on chromosome 4 was found to be linked to white matter hyperintensity (WMH). This sets the stage for narrowing the search to find the gene or genes responsible for inter-individual variability in WMH. This is important to public health because WMH has been shown to associate with cognitive impairment. At present, there is no proven therapy for preventing WMH beyond control of conventional risk factors.
  • Heritability estimates have been generated for many ultrasonographically defined characteristics of carotid arterial disease including strain, distensibility, and stiffness, providing additional scientific rationale for genomic studies using carotid characteristics as a quantitative trait.
  • Phosphodiesterase 4D has been shown to be a risk factor for ischemic stroke in several studies, including 2 population-based studies. PDE4D may be a novel target for stroke prevention.

Unresolved Challenges

The major lesson from the past five years of human genetics research is the central importance of obtaining large numbers of DNA samples from meticulously characterized subjects. This principal applies in particular to the field of stroke genetics because of the disease's heterogeneity and the consequent need for subgroup analysis. The major challenge in stroke genetics is therefore to facilitate the collection and systematic phenotyping of large patient populations.

One of the major challenges towards this end is the identification and definition of heritable stroke phenotypes. Quantitative intermediate phenotypes, such as intima-media thickness (IMT), white matter disease, vascular inflammation, hypercoagulability, or vasomotor reactivity, offer significant potential for future stroke genetic studies. These endophenotypes are easily standardized, which would facilitate multicenter data collection.

In pediatric stroke, there is little information about genetic risk factors. There are 4 known sib-pairs with pediatric stroke, all of whom have known thrombophilic disorders. In moya moya syndrome there are known mutations associated with the vasculopathy as well as several affected families and known haplotypes associated with stroke. The same is true in sickle cell disease where sickle cell disease siblings of stroke children are at greater risk of stroke than unrelated sickle cell patients.

The NINDS DNA Repository has no genetic samples of childhood arterial ischemic stroke. This is a significant deficiency, as mutations may be of great importance in pediatric stroke, which is a non-atherosclerotic stroke model with or without vasculopathy. The genetics of pediatric stroke may be informative for all strokes. It is necessary to obtain clinical, epidemiological, and genetic information in children with pediatric stroke and compare results with an adult cohort, both to understand pediatric and adult stroke.

Genetic research in clinical trials typically has either been relegated to optional translational sub-studies or not done at all. This approach must be radically revised. Genetic risk stratification and pharmacogenomics should be a fundamental design feature of future trials of stroke prevention and treatment. The phase II trial of a FLAP inhibitor for individuals genetically at risk for coronary heart disease, recently reported in JAMA, should serve as a paradigm for future stroke trials. In that study, patients carrying at-risk variants in the FLAP and leukotriene A4 hydrolase genes were randomized to either a FLAP inhibitor (DG-031) or placebo in a crossover design. They found that inflammatory biomarkers of coronary artery disease were suppressed in patients receiving active therapy. Using genetic testing for selecting high-risk or high-responder populations could potentially increase the efficiency and cost-effectiveness of trials by magnifying absolute risk reduction and reducing sample size.

Regulatory restrictions remain an obstacle for conducting retrospective and large multicenter genetic studies. Establishing a uniform approach to the process of collecting and analyzing genetic information, particularly within the context of an NINDS consortium, would help expedite advances in this field.

Priorities and Actions Needed

  • Consensus on the definition, selection, and prioritization of stroke phenotypes, risk factors, and key environmental exposures
  • Collaboration among investigators to collect DNA and well-characterized phenotypic data to rapidly execute large genetics studies in adult and pediatric stroke.
    • Encouragement of DNA sample submission in NINDS-supported clinical trials for pharmacogenomics studies.
    • Recognition of the scientific importance of discovery-driven methodological approaches like whole genome association.

The 2001 Stroke PRG highlighted problems inherent to genetic research of complex medical diseases, including phenotype heterogeneity, lack of standardization in defining endophenotypes, and difficulty in quantifying gene-environment interactions and made recommendations to improve methodology and focus resources.

Looking toward the exciting prospect of applying emerging technologies to whole genome association studies in stroke, there is a strong incentive to develop large-scale collaborations across stroke research centers. Experience with other complex diseases such as diabetes mellitus, rheumatoid arthritis, and myocardial infarction has emphasized the importance of assembling large numbers of DNA samples to detect relatively modest effects on risk, while still limiting false-positive findings via replication. The importance of large samples is magnified still further in the stroke field because the heterogeneity of the disease makes subphenotype analysis an inherent part of any gene search. A solution to this challenge is creation of a Stroke Genetics Study Group, modeled on similar large-scale study groups for other complex diseases. The Stroke Genetics Study Group could effectively build on existing genetics-related programs. Such programs include the Human Genetic Resource Center: DNA and Cell Line Repository (NOT-NS-02-012); the Administrative Supplement to Clinical Studies for Collection of Blood Samples and Data for Repository Banking in Epilepsy, Parkinson's Disease, and Stroke (NOT-NS-03-016); Gene Discovery for Complex Neurological and Neurobehavioral Disorders (PAS-03-092); Microarray Centers for Research on the Nervous System (NS-02-001), and the Administrative Supplements For DNA Microarray Analysis (NOT-NS-04-002).

New Challenges and Opportunities

There has been remarkably little translational stroke genetics research, especially with a focus on gene-environment interactions. One exception is a recent study examining the genetics of small vessel disease. Gould, et al identified mutations in the type IV collagen alpha 1 (COL4A1) gene in mice and human families with porencephaly, and tested the hypothesis that this mutation, in combination with environmental stress, caused hemorrhages in animal and human models. The humans studied were part of a French pedigree with small vessel disease, manifesting with retinal arteriolar tortuosity, leukoencephalopathy, and cerebral hemorrhage. The study demonstrated a predisposition to stress-induced hemorrhage and stroke in carriers of the COL4A1 mutation, although there was wide phenotypic variation. While genetic mediation of stroke susceptibility, evolution, and recovery can be addressed by human linkage, candidate-gene and genome-wide association studies, these approaches will not easily generate answers to questions of gene function.

The use of inbred and congenic rat strains and mouse models in functional genomic research has much to offer in the understanding of the genetic predisposition to stroke.

Answers to the compelling questions of genetic mediation of stroke susceptibility, evolution, and recovery will best be obtained by the complementary methods of linkage, candidate gene studies, and whole-genome association studies. Analytical methods need to be developed to combine genetic and environmental data.

NINDS could act as a catalyst to create a Stroke Genetics Study Group by encouraging investigator-initiated consortia and by removing barriers to cross-institutional collaboration. The mission of this group would be to develop standardized data collection schemes for stroke subphenotypes, assess heritability of the candidate phenotypes, and apply cutting-edge genetic techniques for whole genome association to the pooled, systematically phenotyped subjects. Since the field of stroke genetics is young and the optimal analytical approaches are not yet known, innovation in analytical approaches should be encouraged by maximizing scientific involvement of the participating institutions and investigators and decentralizing the planning and execution of the analyses.

Over the last several years, large epidemiological NIH-funded studies have collected data and DNA on subjects with clinical or subclinical cardiovascular disorders, which may be valuable to investigators studying cerebrovascular disease. These phenotypes include stroke, dementia, cognitive decline, infarcts and leukoaraiosis by MRI, and retinal vascular changes. Several of these data sets are available as "Limited Access Data Sets" through NHLBI (, and at least one, the Framingham Heart Study, provides DNA specimens to qualified investigators free of charge ( In general, however, such resources are poorly catalogued and advertised. Determining existing sources of data would be an important first step in launching a large, collaborative effort in this area. Establishing a structure and process for data storage, access, and publication would need to be done a priori.

There is a growing understanding of the pharmacogenomics of drugs that are commonly used for stroke prevention. An example would be the recent discovery of the role of vitamin K epoxide reductase complex subunit 1 (VKORC1) in warfarin dose determination. These discoveries need to be incorporated into clinical practice. There may be a disincentive for pharmaceutical companies to identify high-responder and low-responder populations through gene testing, thereby creating the opportunity for more focused and restricted use of existing and novel drugs. NIH leadership will therefore be crucial if patient selection for pharmacotherapy is to advance beyond the status quo ante of clinical risk stratification.

The field is poised to make dramatic advances in the next five to ten years. Properly powered whole genome association studies have the potential to identify key stroke susceptibility genes in the U.S. population. The enormous potential of this approach and the considerable resources required to perform these studies require a concerted effort from the stroke community to devise standardized phenotype measures that allow inter-study comparisons and pooling of data. Importantly, investigators wishing to initiate these types of studies should be strongly encouraged to share data on phenotypes, exposures, and genotypes.

An investment in the collection of such data and samples is critical to the long-term success of genetic investigation of stroke; this will not only bear fruit in the application of genome wide SNP typing but also for future genome-scale technologies. In the next period it is likely that cost effective approaches to assess genome-wide allelic expression, epigenetic alterations and ultimately sequence will become available. It is incumbent upon us to ensure that the sample and data sharing infrastructure is in place to take full advantage of these technologies as they arise.

While unbiased genome-wide approaches are just now being used to investigate genetic variability that underlies lifetime risk for disease, it is clear that these data will also provide invaluable information on the role of genetic influences in responses to stroke therapies and neurological recovery. Future studies, which incorporate functional bioassays and neuroimaging, could propel stroke management forward by providing a mechanism for stratifying patients for specific treatments, and identifying new targets for therapy.

Therefore, despite recent NINDS initiatives designed to overcome barriers to genetic stroke research, the time is right to augment current efforts and build the framework of a cohesive, goal-oriented translational stroke genetics consortia. It is imperative that well-designed prospective genetic studies with a priori aims and endophenotypes using a variety of complementary approaches, particularly those focused on pharmacogenomics, be supported in the future.



Co-Chairs: Mark Alberts, Anne Alexandrov, Sharon Ostwald

Members: Mona Bahouth, M.J. Hampel, William Hanson, Edward Jauch, Brett Meyer

NINDS Liaison: Daofen Chen


Stroke consistently ranks in the top 10 of medical diagnoses admitted to U.S. hospitals, necessitating community awareness and interdisciplinary provider preparedness to support optimal patient outcomes. In 1997 and 2002 the National Institute of Neurological Disorders and Stroke (NINDS) of the National Institutes of Health (NIH) recommended improved coordination among the prehospital and acute hospital stroke service sectors aimed at increasing the number of patients eligible to receive acute stroke treatment. These recommendations are in concert with those proposed by the Institute of Medicine (IOM) for establishment of coordinated healthcare systems capable of integrating evidence-based preventive and acute treatment services to enhance public health.

For the purpose of this report, we define "health service initiatives" as, health care delivery systems that span the continuum of care from the acute prehospital environment, to hospital-based services, and on to post-acute rehabilitation, long term care community-based care settings. The past 5-years have seen considerable growth in health services for stroke. We present an overview of key accomplishments within care delivery systems, currently evolving health care models, and provide recommendations for future research to support comprehensive, continuum-based stroke services.

Seminal Scientific Advances

Hyperacute & Acute Stroke Health Services

The past decade has witnessed the dawn of acute stroke treatment with the completion of the successful NINDS rt-PA trial and FDA approval of rt-PA for acute ischemic stroke. A number of significant events affecting the prehospital and hospital sectors were triggered by the commercial release of rt-PA for acute stroke treatment (Table 1). Chief among these accomplishments are the release of guidelines supporting evidence-based acute stroke care, official certification of stroke treatment centers, legislative action endorsing stroke center designation and emergency medical systems transport regulations, use of innovative stroke care delivery models enabling extension of stroke specialists to underserved areas, and hospital reimbursement for treatment with intravenous rt-PA in acute ischemic stroke.

Publication of evidence-based guidelines has laid the foundation for development of formal certification processes that have been implemented by government and accreditation agencies. In particular, the Joint Commission on Accreditation of Healthcare Organizations (JCAHO), Primary Stroke Center certification program has achieved certification of approximately 5% of acute care hospitals nationwide (Table 2). Legislation aimed at improving access and delivery of acute stroke care has been enacted in a number of states, and growing numbers of organized stroke centers capable of delivering byperacute and acute stroke services are emerging across the country.

Table 1: 1996-2006 Acute Stroke Health Services Accomplishments

Top 5 Accomplishments
1. Guidelines for the acute treatment of stroke and stroke center development.
2. Governmental and accrediting agency stroke center certification.
3. Evolving state legislative efforts.
4. Innovative stroke care delivery models: telestroke, electronic-ICU (e-ICU), use of advanced practice nurses (APNs).
5. Hospital reimbursement for intravenous rt-PA.
Other Important Accomplishments

6. Prehospital stroke scales.
7. Emergent prehospital and hospital stroke triage.
8. Recognition of the value of stroke teams and stroke units to patient outcomes.
9. Emergence of successful, model primary and comprehensive stroke programs.
10. Inclusion of stroke guidelines in Advanced Cardiac Life Support.
11. Advanced neuroimaging capabilities to enhance patient selection and treatment.
12. Intraarterial rescue procedures and emerging complimentary methods to enhance intravenous tPA treatment.
13. Cost-effectiveness research supporting thrombolysis.

Table 2: Joint Commission on Accreditation of Healthcare Organizations Certified Primary Stroke Centers (PSCs)

Number of Certified PSCsNumber of States
(includes the District of Columbia)
1 - 311
4 - 66
7 - 96
10 - 307

Note : Florida has the most certified PSCs, driven by state legislative mandate.

The Brain Attack Coalition's (BAC) Guidelines for Comprehensive Stroke Centers were published in 2005 to provide a framework for all-inclusive stroke care. Although no formal certification process currently exists to support these guidelines, it is estimated that 99% of the total US population has access to neurointerventional treatment within a 200-mile radius across 45 states, if emergent transportation systems could provide transport within 6 hours of symptom onset. These centers may be well positioned to offer fully-integrated health services for stroke through regional partnerships with emergency medical services, primary stroke centers and post-acute providers, whose impact, similar to trauma, may serve as a powerful model for effective primordial/primary prevention, treatment, rehabilitation, secondary prevention, and community reintegration.

Incorporation of stroke guidelines into the American Heart Association's Advanced Cardiac Life Support program during the past decade, has fostered dissemination of critical information on stroke risk factors, warning signs, assessment and management to tens of thousands of health care professionals each year. Physician providers are adopting innovative models that incorporate advanced practice nurses (APNs) in the early assessment, and hyperacute/acute treatment of stroke patients, as well as using APNs' skill in system development to expand stroke services across the continuum, but little is known of the safety and effectiveness of APNs as frontline acute stroke providers. Advanced neuroimaging capabilities have enhanced identification of acute stroke patients and potentially salvageable penumbra, as well as identifying patients that may be at increased risk for hemorrhagic transformation, and phase II studies have demonstrated potential benefit of intraarterial, as well as combined intravenous and intraarterial, or ultrasound-enhanced recanalization approaches.

Absence of specialty practitioner availability may account for low rt-PA treatment rates. One third of the U.S. population lives in a rural area, making absence of specialty expertise even more of an issue in these settings. Historically, one method of outreach for these providers has been telephone consultation with stroke trained practitioners. Unfortunately, these techniques limit the amount of information available to the consulting provider (no ability to visualize the patient, engage in verbal and non- verbal informed consent discussions, or visualize integral head CT scan images). Transporting patients to tertiary care facilities, either prior to or after rt-PA administration, may also be helpful in some settings, but, time and distance preclude transfer of all patients.

Though efficacy is still in question, and its assessment is necessary, stroke telemedicine applications have been able to fill many gaps for consulting providers. Numerous telemedicine systems are now in use, and reliability has generally been established. Most telemedicine systems in use today utilize fixed communication lines (ISDN or fiberoptic) to enable point-to-point communications. Newer systems, utilizing web-based or site-independent technologies have been developed and may allow broadened deployment. To fully integrate services, immediate consultant availability is necessary, but currently, practitioners with stroke expertise are not universally available. Freeing up consultants from a fixed base station through use of site-independent telemedicine technologies, and use of APNs to supplement telemedicine response may enable round the clock coverage regardless of geographic location.

While data are limited, the feasibility of critical care-based telemedicine, also called "electronic ICUs" (e-ICUs) has been demonstrated with evidence of improved provider and family satisfaction, as well as improved patient and financial outcomes. Specialty-specific critical care telemedicine targeted at certain patient populations, such as the acute neurology or neurosurgical patient, provides the possibility of intelligent triage and remote specialty consultation in underserved units. Patient populations in whom critical care telemedicine has been investigated include maxillofacial trauma, burns, trauma, high-risk infants, neurology and cardiology. The full promise of critical care telemedicine will be realized with the maturation of the technology and the evolution of standard approaches to licensure, staffing and malpractice. However, evaluation of the application of telemedicine to the ongoing oversight of stroke patients following acute diagnosis and treatment, warrants further investigation given the state of the art.

Post-Acute Stroke Care Services

Among the 700,000 who experience stroke each year, 10% recover almost completely, 25% recover with mild impairments, 40% experience moderate to severe impairments requiring special care, 10% require care in a long term care facility, and 15% die shortly after the stroke. Although 40% of stroke survivors return home with significant impairments and 10% go to nursing homes, the existing five scientific priorities for NINDS do not address issues related to maintaining the long-term functional status of stroke survivors, maintaining compliance with self-management regimes, the impact of stroke-related disabilities on family caregivers, or quality of life for stroke survivors and their family caregivers. These are important areas for research, given that there are almost 5,000,000 stroke survivors living in the United States today.

Researchers who have contacted stroke survivors and their family caregivers after discharge reported that they were frustrated with the lack of information that they received before discharge and continued to have questions during the first 6 months at home. Forster and colleagues conducted a systematic review to examine strategies for providing information to stroke survivors and their caregivers. They concluded. that future work should be directed toward addressing the expressed needs of stroke survivors and their caregivers and identifying appropriate teaching strategies, which can be successfully implemented in clinical practice.

Research on patients with stroke and other common co-morbidities (e.g. hypertension and diabetes) have shown that adherence with self-care management related to taking medications, following dietary and exercise guidelines, smoking cessation, and self monitoring is generally poor, leading to the increased risk of a recurrent disabling stroke or other CV events. Most patients do not comply with medical treatments unless special attention is given to adherence.

Recent research studies have demonstrated that programs that begin early, and are tailored to the individual's educational and skill levels and incorporate self-management strategies, can result in high rates of adherence and successful health outcomes. In addition, Grant and colleagues demonstrated that problem solving skill training for family caregivers can result in greater caregiver preparedness and less depression, as well as greater satisfaction with health care services. Family caregivers are key to stroke survivors being able to remain at home and so are an important target for intervention research, as well.

Key Unresolved Scientific Questions

A lack of public knowledge, and seamless integrated health services throughout much of the nation continues to hamper access to acute stroke services, as well as throughput to rehabilitation and ongoing secondary prevention services necessary to maintain wellness and facilitate return to societal productivity. Recently, the American Stroke Association advocated for the establishment of stroke systems of care to alleviate this problem. According to the ASA, these systems should consist of primordial/primary prevention, community education, emergency medical services, acute and subacute stroke treatment, secondary prevention, and rehabilitation, with each of these phases supported by system-wide continuous quality improvement capable of identifying areas of fragmentation, opportunities for improved coordination and patient throughput, and measurement of the contribution of the system to reduction of stroke burden.

Barriers to implementation of stroke systems include the cost of system organization and maintenance, limited reimbursement for individual institutions and health care providers, local/regional/state public health regulations and/or legislation that may limit effective organization and transport mechanisms, and regional market competition. While intravenous thrombolysis has been shown to be cost effective to society at large, appropriate reimbursement for acute stroke services has only recently become reality, while provider reimbursement for hyperacute care remains elusive.

Research has uncovered better ways to prevent stroke and better treatment for acute stroke, but within the post-acute sector, millions of stroke survivors and their families continue to struggle with the long term impact of stroke on their lives. Hopman and Verner noted that substantial gains in health-related quality of life during inpatient stroke rehabilitation may be followed by equally substantial declines in the 6 months after discharge. Therefore, they recommended the conduct of longitudinal research into the health related quality of life of stroke survivors and their families, as well as a need to ensure that adequate community services and support are available. Sampsa et al, conducted a study on the relationship between functional status and QOL among persons with stroke concluded that various patients make quite different responses to the same disabling condition and that perceived QOL depends on much more than one's level of physical function.

Unresolved scientific questions related to the socio-economic systems barriers described above include discovery of:

1. Methods to fully integrate the health care continuum.
2. Methods to facilitate standardization of evidence-based stroke care across the continuum.
3. Methods to drive widespread community engagement in stroke prevention, response to warning signs, and reintegration/support for stroke survivors and families.

What Needs To Be Done To Address These Questions

Table 3 provides recommendations supporting improvement in continuum-based stroke services. Currently, there is little accountability for the quality of stroke health care services that are delivered in the U.S. While JCAHO certified centers are required to demonstrate use of evidence-based acute stroke care to maintain certification status, this certification remains voluntary with alternative pathways to stroke center designation existing in some states that are not consistently supported by penalties for substandard performance. Additionally, stroke process/outcome indicators are largely ignored by payers. Within the post-acute sector, systems enabling widespread study of health care provider processes and patient outcome are completely lacking, making cross-continuum measurement a challenging endeavor. The opportunity to integrate the stroke health care continuum and facilitate standardization of evidence-based stroke care would be greatly enhanced by methods capable of measuring and making public the end-results of stroke in relation to provider performance.

Acute stroke care providers measure a number of process variables, such as the number of rt-PA treatments provided and the provision of secondary prevention measures. Some data systems are supported by standardized taxonomy and measurement criteria, while others are homegrown databases made up of heterogeneous variables with varying definitions. While certified centers are required to measure standard elements, definitions remain non-uniform and collection of outcome data is rare.

Elements 3 and 4 in Table 3 support recommendations related to data systems that would lend clarity to the understanding of physiologic, psychosocial, economic and system contributors to stroke patient outcomes across the continuum. Ideally, uniform data systems supported by data dictionaries endorsed by key stakeholders would enable pooling of stroke process and outcome data on a large scale enabling a better understanding of system contributors to positive, as well as, untoward outcomes.

Table 3: Recommendations Supporting Improved Continuum-Based Stroke Services

1. Provider and system payment incentives (pay for performance), along with availability of cost-effective technology strategies.
2. Public performance reporting if appropriate severity adjustment (not risk adjustment) measures can be assured.
3. Development of data dictionaries endorsed by key stakeholders, supported by easily adopted, clinically integrated, prospective data storage/shuttling systems.
4. Inclusion of core outcome measures and contributing, as well as confounding, variables that may contribute to long term results.

Additional study of cost-benefit and the overall societal impact of stroke might leverage support for rationale reimbursement schemes among stroke system sectors. Similarly, knowledge of the impact of fully integrated health services on national stroke burden may serve to tear down regulatory and/or legislative barriers to the development and implementation of innovative, amalgamated strategies. Lastly, measurement of population-based impact would enable improved generalization of successful methods to regional systems.

New Research Areas That Have Emerged Since Sprg I

The contribution of certified stroke centers to improved patient and system outcomes has become a new area of research since the emergence of accreditation agency and government stroke center designation. The few studies conducted in this area thus far have been primarily focused on process compliance, while an ability to measure overall societal impact on stroke burden has yet to be demonstrated. Despite this, we expect research in this area to increase significantly in the coming years, as construction of a sound health system infrastructure provides essential support for both the conduct of research and the standardization of practice.

Within the post-acute continuum, health system analyses have yet to emerge, but will become a necessary part of future research as investigators seek discovery of methods capable of supporting integration of acute and post-acute stroke services. Research that has emerged to date in the post-acute sector has focused on the impact of stroke on daily functioning and quality of life.

Priorities Supporting Stroke Health Services Research For The Next Decade

Table 4 lists the research priorities identified by our Health Services Initiatives Workgroup. Although a number of significant advancements have occurred in the area of health services without the benefit of specific direction or significant funding, it is imperative that efforts be focused on the identification of optimal methods for development and refinement of stroke care systems to enable delivery of approved stroke therapies, primary and secondary prevention, and enrollment in clinical research trials.

TABLE 4: Stroke Health Services Research Priorities

1. Understand the contribution of evidence-based, stroke care systems to reduction of stroke burden within the United States population.
2. Identification of methods capable of galvanizing long-term community engagement in the prevention, early recognition, treatment and optimal societal reintegration of stroke.
3. Develop and test the impact of innovative methods of stroke care delivery on access to medical services, cost and resource utilization, and ultimately patient physiologic and psychosocial outcomes.

The overall goal of stroke research and system development is to deliver the best care possible to the greatest number of stroke patients. Therapies with the most generalzability will have the potential for greatest impact. If research solely focuses on sophisticated imaging and therapeutic techniques which can only be performed and delivered at a few select centers, then the vast majority of stroke patients will not benefit from these advances. Similarly, fiscal incentives in the healthcare environment may drive stroke care by reimbursement, instead of best practice. All stroke stakeholders, from healthcare providers to consumers and caregivers, must become engaged in advocating for improved access and seamless health service delivery, provided in an evidence-based framework. Without such a system, testing and implementation of future stroke therapies will continue to be challenged by non-uniform distribution of services.



Co-Chairs: Stephan A. Mayer, Neil Martin, Julian T. Hoff

Members: Dan Hanley, J. Claude Hemphill, William L. Young, E. Sander Connolly, Fernando Vinuela, Jarek Aronowski, Ken Wagner, Paul Vespa, Adnan Qureshi,

NINDS Liaison: Katie Woodbury-Harris


Hemorrhagic stroke -- specifically spontaneous intrcerebral hemorrhage (ICH), aneurysmal subarahnoid hemorrhage (SAH), and brain arteriovenous malformations (AVM) -- comprise altogether only 20% of all cases of stroke, but cause a disproportionate amount of stroke-related death and disability.

Compared to ischemic stroke, hemorrhagic stroke has been relatively neglected with regard to basic science research and clinical trials, but this situation is now rapidly changing. The recent maturation of the field is evidenced by the fact that it is only with this interim 5-year review of the NINDS Stroke Progress Review Group (SPRG) that hemorrhagic stroke has been assigned a separate sub-committee determine research priorities.

The emphasis of the Thrombosis and Hemorrhage section of the original 2001 SPRG was on the need to elucidate the basic science interactions between blood, vessel, and brain. Disruption of the integrity of cerebral vessels is the critical lesion in hemorrhagic stroke, and remains an important focus of emphasis for basic science. Coagulation and thrombosis, the body's main defense against bleeding, has now been demonstrated to lead to exacerbation of brain tissue injury. The interactions of blood, vessel and brain are complex and not fully understood.

Barriers to progress in hemorrhagic stroke research were identified in the 2001 report. These included: (1) Poor interdisciplinary collaboration between researchers in vascular biology, hemostasis, and neurobiology; (2) The paucity of good animal models for many forms of hemorrhagic stroke and the problem of translational relevance of rodent models to the human condition; (3) The discordance between experimental work on cerebral vessels at the molecular level, and underdeveloped methods for biochemical and molecular imaging of human hemorrhagic stroke; and (4) Problems related to clinical trial methodology, in particular the lack of rational preclinical testing and systematic and rigorous phase II dose ranging and feasibility studies. All four of these concerns remain problematic today, and are specifically addressed in the current updated recommendations.



  • Novo-7 ICH Trial (a promising new platform for the emergency treatment of ICH)
  • STICH (will focus surgeons on minimally invasive surgical techniques)
  • ISAT (has established endovascular coiling as an attractive alternative to clipping)
  • Role of thrombin and inflammation in blood-induced brain injury (has opened a new target for intervention)


Major advances in non-traumatic intracerebral hemorrhage (ICH) within the past 5 years have come from two main areas: successful execution of several large clinical trials for acute ICH, and improved understanding of mechanisms of peri-hematoma injury from animal models and clinical research.

Recombinant Activated Factor VII. The rFVIIa study (n=399) confirmed the importance of hematoma expansion on ICH outcome, and demonstrated that this intervention can successfully impact on an intermediary surrogate outcome (CT evidence of hematoma expansion) and a primary clinical outcomes (mortality and functional recovery. The phase III FAST trial will attempt to confirm these findings.

STICH Trial. The Surgical Trial of Intracerebral Hemorrhage (n=1033) demonstrated that a strategy of surgical hematoma evacuation within 72 hours was not superior to initial medical management in patients with supratentorial ICH. A new trial (STICH 2) limited to lobar hemorrhage patients is being planned.

CHANT and SAINT 1. CHANT (n=603) was designed primarily as a safety study, but did not suggest that neuroprotective therapy using a free-radical trapping agent was helpful for ICH. However, SAINT 1 showed a significant reduction in hemorrhagic conversion in AIS patients treated with lytic therapy. These studies have redirected focus on the concept of vascular neuroprotection.

Neuro-hemo-inflammation. A shift away from peri-hematoma ischemia as a major mechanism of ICH-related injury has provided new focus on targets for intervention for ICH-related cellular injury. Animal models suggest that major mechanisms of injury including thrombin-mediated injury to the blood-brain barrier, heme and iron-related cellular toxicity, and inflammation. This, coupled with the rarity of true ischemia from human neuroimaging studies, has changed the direction for clinical targets for ICH treatment.


As is the case with ICH, the most important advances that have impacted on the lives of SAH patients stem from large-scale clinical trials and observational outcomes studies. The advances described below have been dually powered by technology (i.e. GDC coils and angioplasty) as well as the conventional "translational" mechanism (hypothermia, endothelin receptor antagonists)

The ISAT Trial. This International Study of Aneurysm Treatment (funded by the U.K. MRC) was the first to prospectively compare two competing treatment modalities for the repair of ruptured intracranial aneurysms (ICAs). Compared to clipping, coiling resulted in a 7% absolute reduction in the risk of poor outcome at one year. The explanation for this is presumably related to fewer peri- and post-operative complications with coiling. ISAT has helped to support widespread adoption of coiling as a viable primary treatment modality for aneurysms.

ISUIA. The large, observational International Study of Unruptured Intracranial Aneurysms (funded by NIH) has helped to define more precisely the risk of bleeding for unruptured ICAs, as well as the risks of surgery for repair of these lesions. This data had provided important data for patient selection for unruptured ICAs.

IHAST. The Induced Hypothermia for Aneurysm Surgery trial (funded by NIH) found no benefit with intra-operative hypothermia during surgery for ICAs. Since hypothermia entails certain risks, including post-operative wound infection and cardiac morbidity, this trial may lead to improved surgical outcomes by reducing the frequency of this intervention in the target population of IHAST, i.e., good grade SAH patients.

Outcomes Research. Several observational statewide and federal registry-based studies have documented an increase in the likelihood of good outcomes when SAH patients are cared for in high-volume centers hat are able to combine increased experience with specialized resources such as neurological intensive care and interventional neuroradiology. This research has helped push regionalization of care and supports the rationale for the development of Comprehensive Stroke Centers.

Endovascular Therapy for Vasospasm. Angioplasty and the use of intra-arterial vasodilators has become increasingly adopted at high volume centers throughout the world. These interventions have been driven primarily by technological advances, but need to be rigorously evaluated in clinical trials.

CONSCIOUS-1. This recent phase 2 trial (N=400) evaluated clazostentan, an endothelin receptor antagonist, for the prevention of vasospasm after SAH. Although a significant reduction in angiographic vasospasm and cerebral infarction was found, side effects such as hypotension and ARDS were more common, resulting in no benefit in outcome. Driven by a compelling body of basic science research, this trial again demonstrates the difficulty of successfully translating bench research to the bedside.


It is axiomatic that invasive therapy needs to carefully consider the trade-off between natural history and treatment risks. A clearer view is emerging regarding this balance for AVM management. A number of studies have quantified morphological and demographic risk factors associated with future bleeding, and there is a first report of genetic risk markers. Work continues on the hypothesis that the consequences of AVM bleeding are less severe that other forms of ICH, including SAH and ICH. A joint writing group has proposed uniform standards for clinical trial reporting.

Risks of Therapy. There has been improvement in risk estimation for treatment, for both microsurgical resection and radiosurgery. A major advance is a growing awareness of the need to segregate ruptured from unruptured cases in outcome analysis. Those with ruptured AVMs tend to have deficits at presentation and generally improve after surgery. By contrast, patients with unruptured AVMs tend to have normal function at presentation, and are susceptible to worsening with surgery.

Gamma Knife Radiosurgery. A large observational study (N=500) documented that radiosurgery results in significant reductions in subsequent hemorrhage rates in patients presenting with ruptured AVM.

ARUBA. The NIH-funded Randomized Trial of Unruptured Brain Arteriovenous Malformations will test whether functional outcome and the risk of spontaneous AVM rupture at 5 years for best medical therapy is superior to procedural intervention (embolization, microsurgical resection or radiosurgery).



  • Lack of understanding how physiological derangements (i.e. fever, hyperglycemia, vasospasm, edema) modify tissue injury and outcome in the clinical setting
  • Pathophysiology (clinical, molecular, and genetic) of hemorrhage-induced brain injury and vascular injury in human brain tissue
  • Specific, real-world practices that influence patient outcome remain poorly understood.


1. Basic Science

Vascular Injury. Little is known about the interactions of the microvascular compartment and the neuronal/glial compartment during acute hemorrhage. Even less is known about the particular contributions of the brain microvasculature to hemostasis and thrombosis. This includes a significant lack of information about how the endothelium and the glial compartments interact with each other and respond to activation of the coagulation system.

Brain Injury. The factors involved in regulation of cell death following ICH need to be explored further. A limited number of experimental studies have addressed issues related to the mechanisms of cell death triggered by the by-products of blood deposited during the acute phase. Use of genomic and proteomic techniques, as well as cell-type-specific genetically modified mice to identify molecules and pathways contributing to cell injury should help in developing new targets for the treatment of ICH.

Lack of human pathological data. It is recognized that there is a great need for collecting tissue specimens from humans after ICH to allow characterization of the nature of pathological changes in order to 1) determine causes underlying vascular bleeding, 2) elucidate molecular and cellular mechanisms underlying ICH-produced (e.g. mass effect, hematoma toxicity) injury to cells in brain parenchyma, and 3) to provide guidelines in developing an animal model of ICH with pathologies similar to those present in humans.

Lack of relevant animal models. More attention should be paid to discourage researchers from using models that are inexact or less accurately portray morphological and biochemical changes of human ICH. There is an urgent need to develop an experimental model of ICH that would allow for conducting research on regulation of hemostasis/thrombosis. Since the prevention of hematoma growth is recognized as an important target for treatment of acute ICH (Factor VIIa), new treatments targeting hematoma growth and rebleeding may need to be evaluated in the experimental models.

Mechanisms of recovery. Another negelected aspect of iCH research is the development of approaches that can be effectively used to improve functional recovery. Experimental studies should not focus only on the consequences or repair mechanisms of gray and white matter damage. Research should take advantage of experimental models using animal species with more abundant white matter content, similar to that present in humans. Modalities such as motor rehabilitation and progenitor cells-mediated facilitation of recovery should be encouraged in appropriate animal models.

Comorbidity. Studies are needed to include both young and aged animals, as well as both sexes. Cerebrovascular risk factors including hypertension, metabolic syndrome, diabetes, insulin resistance, vascular amyloid angiopathy need to be incorporated into ICH models.

B. Clinical Science

Chronic arterial hypertension is the commonest cause of ICH implying that hemorrhage can be prevented by aggressive medical management. Once ICH has occurred, management shifts to prevention of rebleeding initially, then reduction of the mass effect and toxicity of the hematoma on brain parenchyma.

Acute BP Control. Since most rebleeding occurs within hours after the ictus and is related to hypertension, aggressive reduction of blood pressure may lessen the risk of recurrent hemorrhage, and minimize the severity of peri-hematoma edema. A recent NINDS panel examining research priorities for ICH identified trials evaluating acute (<3 hours) aggressive BP control as the single highest priority for new trials of medical intervention for ICH.

Surgical approaches to ICH. Given the failure of the STICH trial to show a benefit of surgery, it is now recognized that minimally invasive surgical techniques may be a more fruitful approach for hemorrhage evacuation. Removal of intraventricular hemorrhage (IVH) by thrombolytic injection has been shown to be feasible and is the basis for CLEAR IVH trial. Two phase II trials funded by the NIH are also currently underway, examining thrombolytic parenchymal clot aspiration (MISTIE) and endoscopic aspiration with or without thrombolytics (\HEME-Surgery trial).

Critical care interventions. ICH patients often experience a protracted course in the ICU, during which they are subject to many secondary physiological insults, such as fever, hyperglycemia, hypoxia, arterial hypertension, elevated intracranial pressure, non-convulsive seizures, and anemia. Clinical epidemiological studies have increasingly linked these insults to poor clinical outcome. There is a need for studies of specific ICU management strategies that modify or ameliorate these secondary physiological insults.

Pharmacologic strategies targeted at peri-hematoma tissue injury. Basic science research has established that tissue injury in peri-hematoma regions, including BBB disruption and edema formation, and apoptotic and necrotic cell death, is the result of a unique form of "neurohemoinflammation" triggered by thrombin and other coagulation proteins. There is a need to develop and test specific pharmacologic interventions that can minimize these processes and reduce perihematomal injury.

Lack of surrogate endpoints and biomarkers. The lack of validated surrogate markers has been cited as an impediment to clinical trial research and drug development in ICH. Studies are needed to determine whether newer intracranial monitoring techniques (brain tissue oxygen and microdialysis), serum biomarkers (s-100, NSE, axonal neurofilaments), or novel ligand neuro-imaging modalities (MR, PET, SPECT) can provide real time in vivo measurements of acute tissue injury that in turn are predictive of long-term patient outcome.

IRB and waiver of consent issues. Although mechanisms exist for the conduct of emergency clinical trials using a federal waiver of consent, the process has been incocnsistent, arduous and time consuming. As a result the waiver has largely been avoided by investigators and trial sponsors. There is a need to re-examine how patient safety can be optimized in emergency research in a way that actually facilitates, rather than hinders, the investigation of promising new treatments for devastating forms of stroke such as ICH.


1. Basic Science

Evidence from both experimental and clinical studies supports the hypothesis that the brain injury that develops after subarachnoid hemorrhage (SAH) is biphasic. Increasing evidence indicates that acute ischemia develops following the initial bleed and is involved in the pathogenesis of SAH-induced neuronal and vascular injury. At present the detailed cellular and molecular mechanisms that underlie these two brain injury events remain unclear. Despite decades of research investigating the mechanisms that underlie delayed injury (i.e. vasospasm), research investigating the biology of acute brain injury has been relatively neglected until recently.

Mechanisms of acute diffuse brain injury: The majority of SAH patient deaths presently result from the initial bleed. Increases in intracranial pressure and reduced cerebral perfusion are involved in the early development of injury. Studies in animal models and humans have demonstrated early development of apoptosis in endothelial cells. Similarly, protective (cellular survival) molecular pathways (e.g. Akt pathway, impaired NO signaling) have also been shown to be induced after SAH. Our understanding of the molecular mechanisms that underlie acute, diffuse, severe brain injury after poor grade SAH remains incomplete.

Vasospasm. Extracts of red blood cells and especially endothelin-1 and oxyhemoglobin are believed to be active participants in the delayed development of vasospasm following SAH. Recently, it has been proposed that free radicals produced in the blood clots surrounding blood vessels after SAH act on bilirubin and its precursors to produce BOXes (Bilirubin OXidized products). These molecules can cause severe vasoconstriction and are hypothesized to act on vascular smooth muscle cells to produce chronic vasospasm. In addition, the appearance of pro-inflammatory mediators including cytokines, chemokines and leukocytes are likely contributors to vasospasm development. The specific details and relative contribution of these factors and their interactions remain poorly defined.

Need for human pathological studies. Detailed studies (metabolomics, proteomics) of the time course of molecular changes in CSF, brain, and vascular tissue from SAH patients are needed. Over-reliance on extrapolation from animal models is seen as an important limitation in the established "translational" model of clinical research; basic research involving human tissue should receive greater emphasis.

Pathogenesis of aneurysm formation and rupture. The pathophysiological mechanisms that underlie aneurysm formation and rupture remain poorly defined. It seems very likely that genomics and proteomic approaches may play an important role in understanding why vessels weaken and rupture, and possibly may lead to treatments that can stabilize vessels at risk in the future.

2. Clinical Science

Centers of excellence. Preliminary studies indicate that organization of care into comprehensive centers of excellence can improve outcome. Prospective study is needed to determine optimal strategies for physician staffing, imaging and interventional equipment variables (e.g. CT-angiography, bi-planar angio-interventional suites, image guided-surgery) and what center-related variables influence outcomes.

Treatment registries. A comprehensive treatment registry would develop common data acquisition protocols, tools, and resources in order to achieve a comprehensive understanding of the natural history of cerebral aneurysms and determine optimal treatment strategies. This registry would entrain data on various surgical and embolization methods, various types of coils and stents and various neuro-ICU treatment approaches into a comprehensive repository for data assessment and mining. This effort would be similar to the NIH Traumatic Coma Data Bank effort and would be charged with generating publications and evidence-based treatment guidelines and further hypotheses- over the next decade.

Study of aneurysm growth, rupture and re-growth. This should focus on chronic health risk factors, the method of primary occlusion after SAH, and influence of secondary prevention strategies, such as blood pressure control. The integration of genetic, hemodynamic and treatment-related factors need to be analyzed using contemporary imaging and interdisciplinary epidemiologic and computational methods.

Development of improved outcome measures after SAH. SAH is unique among other forms of stroke in its tendency to cause disabling neurocognitive rather than motor and physical disability. Apart from the existing ordinal scale measures of global disability and handicap (i.e. the modified Rankin Scale), there is a need for novel and robust measures that capture more subtle cognitive, emotional, and quality of life outcomes after SAH.

Biomarkers of secondary brain injury. The prospective study of the natural history of secondary injury and cell death, as well as the development of effective biomarkers quantify secondary injury, would be facilitated by the comprehensive registry. Available brain monitoring systems including microdialysis, EEG, MRI, and other means are available but validation of these markers and study of the mechanisms of cell death and cellular regeneration/re-organization are needed.

Vasospasm. Effective means for preventing and treating vasospasm remain unresolved, including medical and interventional strategies aimed at restoring cerebral blood flow, neurochemical manipulation, the use of neuroprotectants such as statins, and other measures.


Many of these challenges are related to not only AVMs but also other related hemorrhagic vascular malformations of the brain including cavernous malformations and dural AV fistulae. For all of these conditions, better detection, screening and treatment decision algorithms are needed.

1. Basic Science

Lack of a Relevant Animal Model. There is no good animal model for brain AVM. Virtually all of the models proposed to-date are extradural fistulas that simulate various hemodynamic or anatomic aberrations. An ideal brain AVM model would require (a) physiologically significant hemodynamic consequences as a result of direct arterio-venous shunting; (b) adjacent brain parenchyma undergoing active angiogenesis and vascular remodeling; and, importantly (c) a syndrome of recurrent, spontaneous ICH.

Lack of Longitudinal Pathophysiological Natural History data. There are inadequate means to study the molecular pathophysiology of the disease in a longitudinal context in intact subjects. Such studies will be possible with improved MR or PET tracer development to label molecular level participants in key mechanistic pathways. Even in the absence of an animal model, such technology may be applicable to the clinical studies.

Human Tissue Studies. Studying resected lesional tissue samples, modest success has been made in identifying angiogenic and inflammatory pathways that are active at the time of treatment. However, these observations may not reflect mechanisms for AVM etiology, formation, growth and symptom generation, which remain largely obscure at present. Human tissue studies have been underemphasized to date.

2. Clinical Science

Mordibity Predictors. Better understanding of patient and operator dependent predictors of morbidity for both natural history and treatment response are needed. There is a natural disparity in outcomes across different sites. Better understanding of variables that influence outcome can take what is learned from highly experienced centers and quantitatively translate that experience to all sites with the development of large patient registries that encompass a diverse network of institutions. At present only a fraction of available case material reaches the scientific literature at the present time, hampering inferences that can be applied to clinical practice.

Human Brain imaging and Biomarkers. Improved brain imaging techniques are needed to assess hemodynamics and the functional eloquence of brain regions involved in the disease process. Biomarkers (i.e. SNPs, inflammatory mediators) are needed that are associated in a predictive capacity to natural history and treatment outcomes. Also critical is the need to develop RNA transcript expression or protein biomarkers in biological fluids and circulating cells. Robust biomarker development in the study of AVMs is particularly important because current risk factors are based on AVM angioarchitecture, and are not well standardized, e.g., lesion size or categorical, e.g., venous drainage, eloquence, prior hemorrhage.



  • Realization that patient outcomes are strongly influenced by center-specific characteristics (i.e. high volume centers).
  • Appreciation of the role of therapeutic nihilism in determining patient outcomes.
  • Role of rapidly evolving technological advances (i.e. endovascular, ICU monitoring and interventions) in driving patient management.

It has become increasingly clear that many "real world" variables that influence the process of care - specifically who is caring for the patient, how aggressively is care pursued, and how are various technological innovations being employed - exert a powerful influence on the lives and outcomes of patients with hemorrhagic stroke. Research in this area has escaped the traditional focus on basic science and clinical trials. There is a need for clinical epidemiologic research that provides a better understanding of how these "real world' variables affect outcome.



  • Investigate the impact of acute physiological management strategies (i.e. BP, temperature, glucose, seizures, osmolality, oxygen, vasospasm) on outcome
  • Create a hemorrhagic stroke human tissue bank for basic science research focusing on blood-neuron-vessel interactions and vascular lesion formation.
  • Develop better animal models that more closely resemble the human condition of hemorrhage and hemorrhagic transformation of ischemic stroke.
  • Develop comprehensive, internet based, national outcome registries for ICH, SAH, and AVM.


Several specific steps can be taken to advance our understanding of how to better treat ICH and improve clinical outcomes. Basic science research focusing on the mechanisms of cell death and tissue injury is needed. Whenever possible this research should involve human brain tissue analysis. Relevant dynamic animal models of acute ICH are needed. Clinical trials of hyperacute (<3 hours) emergency BP reduction and other physiologic interventions are needed. Surgical trials should focus on minimally invasive acute clot evacuation techniques. Large-scale ICH patient registries may help to develop valid surrogate endpoints for ICH research, improved measures of outcome, and a better understanding of center-related and technologically-driven practice variations that influence outcome. The establishment of a nationwide, federally sanctioned IRB to facilitate the safe and ethical conduct of acute stroke research would particularly benefit victims of ICH, who lack capacity to a disproportionate degree. This would particularly facilitate pre-hospital trials, which may soon be feasible with the advent of remote ambulance-based brain imaging.


Basic science research focused on the mechanisms of acute diffuse tissue and neurovascular injury after SAH are needed, in addition to continued research focusing on the molecular mechanisms of vasospasm. Again, more emphasis on human tissue basic science research is needed. Health care outcomes research to better understand center-related variables that influence outcome is needed. This would also facilitate (1) the collection of data on the safety of aneurysm treatment, (2) better understanding of the natural history of aneurysm growth and rupture, (3) the validation of novel and robust outcome measures, and (4) the validation of surrogate markers of acute and secondary brain injury.


Nationwide AVM detection and outcomes databases or registries would help researchers gain knowledge and understanding of this disease from the vast majority of patients who currently cared for in isolation. Improved disease-specific clinical trial methods (e.g., adjustment for severity, outcome adjudication, biomarker development) are needed. Central repositories of genetic and other biological specimens are needed in order to facilitate biomarker development and basic science research. Knowledge gained from studies of AVM tissues and genetics needs to be used to develop medical therapies for prevention of AVM rupture, such as MMP-9 inhibitors or other anti-inflammatory or anti-angiogenic approaches. Progress has also continued in the development of novel endovascular and radiosurgical approaches to treating AVMs, which can be tracked in large observational databases.



Co-chairs: Gregory J. del Zoppo, David J.Pinsky

Members: Shaun Caughlin, Bruce Coull, Paula Dore-Duffy, John Hallenbeck, John Harlan, Eng Lo, Joseph Loscalzo, James H. Morrissey, Maiken Nedergaard, Ed Plow, Jordan Pober, Bruce Ranso, Marc Simard, Stella Tsirka

NINDS Liaison: Thomas Jacobs, Michael Moskowitz

The role of the cerebrovascular endothelium in response to focal ischemia was examined. It was concluded by the participants that the cerebrovascular endothelium must be regarded in the context of its proximity to astrocytes beyond the basal lamina matrix within the neurovascular unit, and their known interactions during development, normal function, and responses to diseases targetted to the neurovascular unit. This is particularly relevant for focal cerebral ischemia. Because the group of participants now included experts in vascular biology in non-neurological settings, there is an additional "flavor" of how what is known about the endothelium in those settings could be "translated" to the cerebral vasculature. Nonetheless, the group unanimously supported the text and conclusions of the section on "CNS Thrombosis and Hemorrhage" in the Report of the Stroke Progress review Group (April 2002) as being current and still forward-looking. We extend those observations and provide further suggestions on,


There have been a number of advances in the last five years are now central to our understanding of hemostasis and endothelial cell responses. These advances occurred both in the specific context of ischemic CNS injury and in unrelated areas which can be applied to the area of stroke. The tissue injury from ischemic stroke involves an interplay in neurons and their supporting cells between the countervailing forces of thrombosis and anti-coagulation, inflammation and pacification, and internal and external cell death and survival signals. Broad conceptual advances have been made in the area of understanding the tissue- and context-specificity of vascular responses. That is, similar stimuli can evoke completely different responses in vascular cells depending on their geographic context (i.e., surrounding tissue), their location within the vasculature (i.e. large arteries, arterioles, capillaries, venous elements, and lymphatics) and their environmental context (i.e., the ambient oxygen, glucose, redox milieu). A specific area of advancement and excitement has been in the area of cellular regeneration, of neurons and their supporting cells, and of endothelial cells. Pleuripotential precursor cells, at various states of differentiation are now believed to reside within the CNS as well as within bone marrow pools where they could assist in the reparative response to ischemic tissue injury in the central nervous system. There is currently intense interest in the receptor-ligand interactions and signaling mechanisms responsible for their recruitment and proliferation in the context of stroke. In the area of thrombosis, particularly germane to the topic of ischemic stroke, there have been advances in understanding of protease cascades and ectoenzymes which participate in vascular homeostasis (such as ADAMTS-13, metalloproteases, and ectonucleotidases). This emphasizes the development of antithrombotic agents which target proteases. The participation of protease activated receptors in neurogenic inflammation has been an important conceptual advance. The role of endothelial and platelet microparticles, as mediators (or suppressors) of coagulation and inflammation and their ability to signal downstream sites, is an area of current intense interest in the context of ischemic and thrombotic diatheses.

There have been advances specific to the CNS as well. These include a greater acceptance of the "neurovascular unit" as a fundamental conceptual basis for understanding vascular regulation and dysregulation in the context of ischemic central nervous system injury. There is a growing recognition of the differential susceptibility of gray and white matter to ischemic insult, the causes of which are currently being elucidated. Pathways by which cells of the CNS die have also been a topic of continued interest in the last five years, exceedingly important in the study of the pathophysiology of stroke. Increased expression of matrix proteases, involved potentially in reparative responses of the CNS to ischemic injury, can also trigger devastating intracerebral hemorrhage. The role of pericytes in the CNS represents a relatively new area of focus. Novel imaging modalities relevant to stroke are emerging and a subject of great interest, both for their ability to provide therapeutic and prognostic information, but also for their potential ability to influence treatment decisions. The role of spreading depression as a contributing factor to post-ischemic neuronal injury in the setting of stroke has received continued attention.

In summary, major scientific advances of the last five years, which have had a positive impact on research include:

1. the conceptualization of ther neurovascular unit and growing scientific support for the concept based on work on its component parts and intercellular signaling processes,

2. the evidence for tissue and context specificity of vascular reponses to ischemia,

3. progress in understanding the roles of proteases and their signaling cascades, and

4. evidence for roles of endothelial cell and platelet microparticles and their participation in inflammation and coagulation.


Originally this section integrated cerebral microvascular properties with neuronal and glial activation. This is the continuing direction of scientific work.

Among the still unresolved targets and the directions for further study is the general need to look at interventions in the context of the time before, during, and after ischemic injury. Temporal factors, while useful for correlational judgments, are still important for defining the responses of the neurovascular unit elements to one another. More specifically, within the neurovascular unit the function of the endothelium is still unexplained. Among the undefined features of the cerebral endothelial lining are i) the nature of the heterogeneity of endothelial cell properties at rest and their responses to focal ischemia, and ii) the specific differences in endothelial cell properties between white matter and gray matter. The roles of the endothelium in the energy demands of the CNS tissue, under normotensive or hypertensive conditions and with age, are likely to provide clinically relevant information, and to help establish appropriate model systems for the future of ischemic stroke research. Furthermore, there has been insufficient work applied to the reparative and angiogenic phases of the cerebral microvasculature after ischemic injury (or any type of injury). These will need new approaches. Among these will be the potential roles of pericytes, adult stem cells, and endothelial transport capabilities in neovascularization. In particular, focus on the permeability barrier properties of the endothelium as a therapeutic target is needed.

The cerebral microvasculature has been considered the only conduit of fluid transit in the CNS. However, there is little information or understanding of the characteristics of the CNS lymphatics, which might have been mistaken for microvessels.

Other important directions should be explored. There is a continued need to extend our understanding of the hemostatic system (both platelet-dependent events and coagulation) within the microvascular compartment at rest and during ischemia. For instance, the expression of tissue factor (TF), its inhibitors, and other modulators of hemostasis within the neurovascular unit and perivascular neuropil should be explored. This could include the use of knockout mouse preparations in models of ischemic stroke for specific questions. This direction is intended to test the hypothesis that CNS-specific hemostasis exists. In cerebral microvessels, pericytes may serve as multipotential stem cells, particularly under conditions of stimulation.

There is a particular need i) to further understand the structure, function, and differential responses of astrocyte end-feet of gray matter and white matter, and ii) to initiate an understanding of the proposed cross-talk between endothelium and astrocytes within the neurovascular unit.

Further exploration of the immune responses of the neurovascular unit in focal cerebral ischemia is needed. This includes the endogenous inflammatory response by astrocytes and microglia as initiated by ischemia. It extends to the heterogeneity of dendritic cell responses and microglial cell responses, and the ability to support immune-modulation to reduce injury.

There is little known about the impact of environmental factors on microvascular responses within the neurovascular unit, especially during focal ischemia. For instance, responses of the vascular beds in gray matter vs white matter to hypertension and age are likely to be highly relevant to the outcomes of ischemic injury, and can be studied in small animal models subject to elevated blood pressure and/or age. Genetically modified mice could be developed to explore/define the microvascular endothelial cell, astrocyte, and neuronal responses to ischemic injury. The use of new micro-imaging techniques is encouraged for concomitant bioenergetic studies.

The temporal context of ischemic injury and the repair of the elements of the neurovascular unit have not been examined as a basis to begin exploring long-term outcome and recovery following focal cerebral ischemia.

In summary, continuing important unresolved questions from the last five years include:

1. understanding the functional interactions between the endothelium and astrocytes within the neurovascular unit, and their modulation by the actions of other cells,

2. the nature of the interactions between the endothelium and the hemostatic system within the cerebral vasculature,

3. the impact of the location and environment to differential responses of the neurovascular unit to focal ischemia, and

4. the roles of the inflammatory responses to ischemia, both intrinsic to and extrinsic to the brain.


There are several critical needs for which NINDS support could be directed. There is a great need to understand the temporal relation of events surrounding stroke, to better understand which observed features are causes and which are consequences of ischemic tissue injury in the CNS. This is particularly important because of the narrow time window of opportunity for certain pharmacologic treatments (such as rt-PA), and the emerging evidence that treatments which are good at one time may be harmful at others. Early on, a treatment which mitigates inflammation may be beneficial (e.g. TNF antagonists), whereas at later times it might impair the reparative response. Importantly, inflammation itself is a great challenge in the stroke arena, as it is not yet proven, despite significant pre-clinical data, that modulating the inflammatory response can improve stroke outcomes. An improved understanding of the temporal sequence of pathophysiological events could help enable delivery of the appropriate therapy at the appropriate time.

Another critical need is to understand the role of disease modifiers, such as age, hypertension, and other modifiers of the cerebral vascular bed, neurons, and their supporting cells in the context of stroke. Recognized complications of stroke such as intracerebral hemorrhage are common in the elderly, or in hypertensive subjects, yet the mechanisms driving this deleterious age response are not well understood. Similarly, hypertension can increases susceptibility to ischemic injury for reasons that remain to be elucidated.

Despite a wealth of pre-clinical data, the majority of clinical trials of new agents for the treatment of stroke have failed. Although each failure can be rationalized by a fault of experimental design, the larger view is that so far animal models have not adequately predicted clinical trial results. This observation highlights the need for improved animal models of stroke (perhaps large animal models), as well as new reductionist models. Proposed models include cellular and organ culture (brain slice) models which could better mirror the interactions of the endothelium, neurons, support cells, and leukocytes in vivo. Hence, both integrative and reductionist approaches are needed to increase the relevance of treatments and to improve the design of clinical trials of new potential therapeutic agents for stroke.

Given the paucity of existing stroke therapies, there is a need to employ sophisticated genetic, epigenetic, and proteomic approaches to new target identification for the development of new therapeutic agents for the treatment of stroke. Several potential areas for which further study is needed include i) the role of specific protease inhibitors, ii) the possibility of targeting endothelial permeability or transport capability, or iii) cross-talk on both sides of the cerebral vessel (e.g., astrocytes inform endothelium, and vice versa).

In summary, to answer these unresolved issues there is need to:

1. clearly define the temporal relationships of the responses of endothelial cells (amd the neurovascular unit) to ischemia (e.g. as pre-ischemia, immediate post-ischemia, and reparative phases),

2. understand potential disease modifiers in relation to tissue susceptibility to focal ischemia an to intracerebral hemorrhage (e.g. hypertension, age, diabetes mellitus), and

3. develop and utilize selective models for studying stroke using both reductionist and integrative approaches (e.g. cells, organ tissue culture, and animal models).


Research into the responses of the cerebral vascular components, and the neurovascular unit in particular, that have emerged since the publication of the Stroke Progress Review Group report of 2002 clearly reflect the initially recommended directions of enquiry.

New areas of fruitful research in vascular biology could apply to the neurovascular unit. These include the role(s) of adult stem cells in the cerebral microvascular responses to ischemic injury. The finding that pericytes as multipotential stem cells migrate towards sites of injury and could participate in injury response highlight these cells as potential regulators of microvessel activation and neogenesis.

Furthermore, understanding endothelial cell responses to ischemia which mediate the increases in permeability is slowly growing, and requires focus on matrix adhesion and local cell-cell interactions.

Cross-talk among cells within the neurovascular unit is a concept of growing significance.

The possibility that hemostatic proteins could behave differently within the neuropil than the vasculature has emerged. This is highlighted by apparent differences in the responses of microvessels a nd neurons to rt-PA in cell culture and small animal models of cerebral injury.

Regarding the cerebral endothelium, insights in the field of vascular biology from a number of non-cerebral venues have increased understanding of potential targets in the cerebral microvasculature. It is now clear that endothelial cell "activation" differs with the specific stimulus, and that the responses are likely to be tissue specific. Receptor expression is controlled by location in the organ and what previous and prevailing conditions existed. It is likely that this "context-dependency" is also relevant to the brain. Furthermore, endothelial cell heterogeneity, as it reflects different gene expression patterns (at both transcriptional and translational levels), is known. There is growing evidence that this occurs in cerebral responses as well. The known intersection between coagulation and inflammation also applies to the cerebral responses to ischemia. This has been confirmed and gives support for this general interaction of the two response systems. It also points out our still woefully insufficient knowledge about hemostatic mechanisms in cerebral vasculature. It is likely that as knowledge of the neurovascular unit grows and the microvascular responses in the context of focal ischemia, that specific handling of hemostasis will emerge.

Specifically for the CNS, several areas of successful research understanding have appeared. Extended knowledge about the acute and very early events following the onset of focal ischemia and their impact on microvascular responses has developed. These include further understanding of the fate of the vascular matrix environment and cellular adhesion receptors. It is clear that both in vivo and in vitro the relative sensitivity of cells within the neurovascular unit is: neuron > oligodendroglia > astrocytes > endothelium. Endothelial cells are relatively resistant to ischemic injury in vivo, and this is reflected in cell culture.

Extended knowledge of the permeability barrier responses of the cerebral microvascular endothelium has provided opportunity to consider strategies for its regulation in vivo by interventions.

The development and use of imaging techniques, particularly multi-photon imaging, has allowed visualization of the cortical vasculature in real-time and responses of neuronal and microvasculature to select stimuli. In addition, it provides real-time evidence of astrocyte activation and microvascular responses. This technique will be invaluable for the examination of early responses to focal ischemia.

There is growing evidence for the participation of protease families in the early evolution of ischemic injury. These include MMPs, serine proteases, and cysteine proteases. This knowledge is still new, and requires significant development. Parallel insight into the actions of ADAMTS-13 and highlights the value of this research. In addition, progress in understanding the role and processes managed by cell surface proteases (e.g. u-PA, t-PA, and others) is likely to translate into new information with therapeutic value for the microvascular responses to ischemia.

Protease-activable receptors (PARs) are employed in many cell activation processes, and are important in thrombin-mediated cell responses. It is now recognized that PARs can participate in CNS inflammation and in monocyte/microglial cell responses. Whether and how they appear in response to focal ischemia is not yet known. It is likely that they could participate in vascular, non-vascular neuronal, and coagulation system responses, perhaps linking some of these effects.

A number of exciting directions of research have emerged in this area since the original report of the Stroke Progress Review Group that illicit attention, and are likely to advance our understanding and the opportunities for developing new treatments:

1. understanding cross-talk among cells within the neurovascular unit,

2. understanding the development and maintenance of the microvascular permeability barrier, and its potential as a therapeutic target,

3. understanding hemostasis specific for the CNS,

4. understanding proteases and their receptors and ligands in the context of focal cerebral; ischemia and post-ischemic inflammation and hemorrhagic transformation, and

5. understanding role(s) of multipotent cells in post-ischemic vascular and neuronal repair in the CNS.


All of these research areas existed at the time of the publication of the original SPRG report. This listing focuses on the cerebral microvasculature and the neurovascular unit, in particular. However, it must also be emphasized that the connection of the proposed responses within the neurovascular unit to neuronal responses is still in the early days.



Co-Chairs: David A. Greenberg, Eng H. Lo, Midori A. Yenari

Members: Dale Bredesen, Jun Chen, Paul Dore-Duffy, Gary Fiskum, Rona Giffard, Peter Stys, Michael Tymianski, Justin Zivin, Berislav Zlokovic,

NINDS Liaison: Eugene Golanov

Advances Post-SPRG-2002

Advances in science tend to be incremental. But since the 2002 Stroke PRG, several interesting discoveries deserve discussion. From a clinical perspective, improvements in clinical trial design, wider use of thrombolytic therapy, elucidation of stroke risk factors, advances in neuroimaging, and expansion of the SPOTRIAS network by NINDS are among the highlights. Most recently, intriguing data from the SAINT-1 trial suggest that a nitrone-based radical spin trap may have beneficial effects in acute ischemic stroke. In this PRG section, however, we will primarily focus on basic science and translational aspects of neurocerebrovascular pathophysiology.

From a neuronal perspective, new data suggest the importance of cell death pathways beyond the traditional NMDA/AMPA glutamate receptor models. Transient receptor potential (TRPM) channels were identified that are activated by free radicals, thus providing cross-talk between oxidative stress and excitotoxicity. Cleavage of sodium-calcium exchangers and other membrane calcium pumps by caspases and calpains provide crosstalk between apoptosis and necrosis. Dysregulation of acid-sensing channels may mediate delayed ionic imbalance beyond glutamate receptors alone. And gap junctions and hemichannels may also contribute to calcium imbalance in stressed neurons.

From a glial perspective, a wealth of new findings suggest that multiple pathologic pathways are triggered in astrocytes, microglia, oligodendroglia, pericytes, and other cell types. Emerging data on aquaporin, connexin, gap junction, and SUR1-regulated NC-Ca-ATP families of astrocytic channels may provide new targets for ameliorating cell swelling and brain edema. New imaging, transgenic and pharmacologic studies now suggest that microglia are highly dynamic and may possess both beneficial and detrimental effects. Oligodendrocytes comprise a critically understudied yet highly vulnerable brain cell population. New data suggest that novel NMDA receptor subunits render oligodendrocytes susceptible to excitotoxicity in soma and distal processes. Detailed knowledge about pericytes is lacking, but emerging data suggest key roles in neurovascular homeostasis and perhaps even cell replacement.

From an endothelial perspective, emerging data are beginning to overturn the dogma that the cerebrovasculature is resistant to stroke. Reverse coupling between neurons and hemodynamic function in post-stroke brain may contribute to the collapsing penumbra. Repair of endothelial function may precede neuronal recovery. Advances in protein transduction technology may allow improved methods to deliver stroke therapeutics into brain. Another major advance may be the notion that angiogenesis is tightly coupled to neurogenesis, thus providing an avenue of cell-cell signaling within the neurovascular unit. And finally, within the bloodstream itself, the biologic potential of circulating endothelial progenitors is beginning to be appreciated.

In the face of these "cell-specific" advances in stroke science, comes the challenge of how to tie it all together. A major advance triggered in part by the 2002 Stroke PRG is the widespread diffusion of the neurovascular unit concept in the field. This has stimulated an explosion of papers, symposiums and research efforts and collaborations across multiple disciplines. There is now an increasing emphasis on the dissection of cell-cell signaling instead of a solitary focus on intra-neuronal pathways alone. Several groups have also taken advantage of the new tools of genomics and proteomics to begin characterization of the systemic response to stroke.

From an administrative perspective, it is increasingly recognized that collaboration is the key. Based in part on recommendations from the 2002 PRG, multiple consortiums have been put into place by the NIH. The Pre-Clinical Stroke Model consortia effort may provide a critical translational counterpart for the clinical SPOTRIAS network. Several RFAs that focus on the integrative biology of the neurovascular unit have been initiated. The U01, R21 and SBIR/STTR project grants provide an avenue for applied stroke research in terms of developing drugs and technology. All of these should continue to be supported. Ultimately, however, it should be emphasized that the investigator-initiated R01 research project has worked well in terms of advancing the basic science and providing a pipeline for therapeutic targets and ideas.

Unresolved Questions

The new emphasis on the neurovascular unit compels us to broaden our focus beyond the solitary neuron to a more integrative approach where we dissect responses of all cell types in brain. What are the molecular mediators of cell-cell signaling between neuronal, glial and vascular compartments in brain? How do these interactions underlie function in normal brain? How do they become uncoupled during stroke? Are there unrecognized regional variations in gray versus white matter, or neocortex versus subcortical structures? How do they mediate tissue remodeling during stroke recovery? Are remote responses "far way" from the central territories of stroke important? How are these neurovascular signals linked to clinically measurable markers? For example, the diffusion-perfusion mismatch is increasingly used as an indirect surrogate marker for a "treatable" penumbra. But what does the biophysical MRI signal of diffusion-perfusion mismatch truly mean in terms of actual underlying pathophysiology at the cellular and molecular levels?

Much progress has been made over the past 5 years since the last PRG refining the mechanisms of excitotoxicity, oxidative stress, inflammation, and programmed (apoptotic and necrotic) cell death. However, stroke is a complex disease, so a focus on singular targets in isolated pathways is unlikely to work. A major question now is how to tie it all together. How are these pathways linked? What are the mediators of crosstalk and redundancy? Can we more effectively interrupt progressing stroke pathophysiology by targeting these crosstalk signals? Can we learn how to manipulate them from dissecting endogenous mechanisms of cell survival from preconditioning and tolerance paradigms?

Many mediators and signaling molecules in stroke have dual roles. For example, overactivation of NMDA receptors is clearly excitotoxic, but subtle homeostatic NMDA signaling is necessary for neuronal survival and plasticity. Microglial activation obviously amplifies inflammatory tissue damage, but under some conditions these cells can also secrete beneficial neurotrophic factors. Another example is the recent discovery that whereas dysregulated matrix metalloproteases underlie neurovascular injury during the acute phase of stroke, a secondary activation of these matrix proteases mediates neurovascular remodeling and cell migration during delayed stages of functional recovery. Even free radicals may have beneficial roles in terms of cell signaling under normal conditions so that a broad untitrated suppression of free radicals may have significant side effects. Where is the transition point between acute damage and beneficial remodeling? How can we titrate our therapies to suppress abnormal signals while not perturbing functional homeostasis? Finding this critical balance may be the key.

Beyond a focus on acute treatments, questions now arise as to whether we can "regrow" brain. New exciting data converging from molecular, cellular, developmental and evolutionary biology strongly suggest that there are significant overlaps between neurogenesis, gliogenesis and angiogenesis. What are the signals that mediate the coordinated response of the entire neurovascular unit as the brain attempts to remodel after injury?

Even if we fully dissect all the complex mediators and pathophysiologic signals in stroke using experimental systems, we may still need to understand how these translate into a heterogenous clinical population. What are the effects of age, gender, race, and lifestyle? For example, the two genes PDE4 and ALOX5AP have been identified as risk factors. But how they fit into our current models of stroke pathophysiology needs to clarified.

Ultimately, the question still remains as to how we can better translate data from molecular, cellular, tissue, and whole animal models into clinically viable targets and therapies. What is the best and longest time-to-treatment window in the best case scenario? The previous PRG seemed to suggest that 3 hrs was the cut-off. This may not be true, depending on the signaling pathways being targeted. Are there critical species differences that prevent us from effectively interpreting our animal model data? A more detailed understanding of species differences in excitotoxicity, oxidative stress, inflammation, and programmed cell death mechanisms is required.

Barriers, Opportunities and Priorities

Stroke is a very challenging disease. It is not only a neurodegenerative disorder but also a medical emergency, so timeframes for diagnosis, decision-making, and acute therapy are compressed. Stroke is not a purely neuronal disorder. It affects all cells in the neurovascular unit, comprising glia and vasculature as well. So therapies must rescue multiple cell types with diverse and complex phenotypes. Although we have made significant progress, several barriers may yet remain.

  • It is critical to investigate all components of the entire neurovascular unit. But the scientific disciplines and expertise are disparate, and collaborative barriers may still exist.
  • It was recognized at the 2002 PRG that our models do not replicate all aspects of human stroke. Translational barriers between experimental models and clinical applications are still present. Standardization of techniques and data collection remain a problem.
  • Advances in genomics and proteomics are exciting. But databases and tissue banks for experimental models are not widely available.
  • The R01 mechanism for funding and encouraging basic research is very strong. Resources to support clinical research via SPOTRIAS and other structures is beginning to come into place. However, there is still a danger that translational research that lies between the two ends of the spectrum may fall between the cracks.
Nevertheless, many opportunities exist.
  • Genomic and proteomic tools are becoming more sophisticated yet more user-friendly. These tools must be leveraged for stroke research.
  • In vivo imaging tools are being developed at an astounding rate. Combination of these tools with transgenic and pharmacologic models should allow us to dissect cell-cell signaling in the neurovascular unit in real time.
  • The increasing collaborations between NINDS and other institutes at NIH should provide powerful opportunities to link stroke science with advances from Alzheimer's disease and other CNS disorders.
  • The new Pre-Clinical Consortium should allow us to standardize models and techniques in parallel with the clinical SPOTRIAS network.
Based on these unresolved questions, barriers, and opportunities, several priorities may be emphasized.
  • The whole animal model is not a human stroke. Instead, we need to create a series of linked platforms spanning the range from in vitro to cell to tissue to in vivo systems. Experimental consortiums should be developed to stimulate collaborations and standardization. Shared databases and tissue banks should help.
  • Experimental readouts tend to be focused on cell death. We need to develop assays that also measure functional readouts within all neuronal, glial and vascular components of our linked model systems.
  • White matter remains relatively understudied. The modular construct of the neurovascular unit comprising microvessel endothelium, pericyte, matrix, astrocytic endfeet, and associated neuron is rapidly becoming a "standard model". A similar construct is needed for white matter to investigate vascular, glial and axonal interactions.
  • Many therapeutic targets may have dual roles in acute injury as well as delayed recovery. We need to understand how these roles shift and balance during stroke progression. What is the transition point between acute injury and remodeling? A better understanding of these transitions may allow us to achieve acute protection while maximizing recovery strategies.
  • Over the past 5 years, tremendous progress has been achieved in developing quantifiable biomarkers including neuroimaging, genetic and blood markers. In order to truly exploit the power of these biomarkers, we need to relate them to the underlying molecular and cellular pathophysiology.
  • Advances in tissue engineering may offer the exciting opportunity to repair brain. Neural cell replacement, glial modifying, and vascular scaffolding bioengineering efforts should be encouraged.



Co-Chairs: Greg Albers, Chelsea Kidwell, Steven Warach

Members: Andrei Alexandrov, Jeff Alger, Steven Cramer, Colin Derdeyn, John Detre, Walter Koroshetz, Randy Marshall, Michael Moseley, Bill Powers, Howard Rowley, Gregory A. Sorensen, Christine Wijman

NINDS Liaison: Deofen Chen

Introduction and Seminal Scientific Advances Made Since 2001

Substantive advances in the field of neuroimaging have continued over the past 5 years since the publication of the first Stroke Progress Review Group in 2001. These have included new developments in stroke imaging diagnostic markers, understanding of stroke pathophysiology. and advances in therapeutic approaches employing neuroimaging techniques for selection, monitoring, and testing of therapeutic interventions, as well as general advances in imaging technologies and implementation.

Neuroimaging for selection, monitoring, and testing of therapeutic interventions

STOP-2 trial: The STOP trials proved the utility of transcranial Doppler ultrasonography (TCD) screening in children with sickle cell anemia for primary prevention via prophylactic transfusion therapy in those patients found to be at high risk by TCD criteria. The more recent STOP 2 trial was a randomized trial demonstrating that discontinuation of transfusion for the prevention of stroke in children whose blood velocities have normalized results in a high rate of reversion to abnormal blood-flow velocities on TCD and stroke.

CLOTBUST trial: Randomized controlled trial of continuous 2-MHz TCD in combination with t-PA found that complete recanalization or dramatic clinical recovery within two hours after the administration of a t-PA bolus was augmented relative to control (tPA without continuous TCD) and trends toward increased clinical recovery.

NOVO-7 intracerebral hemorrhage trial: Treatment with recombinant activated factor VII within four hours after the onset of intracerebral hemorrhage reduced the growth of the hematoma on CT relative to pre-treatment hematoma volume, associated with reduced mortality and improved functional outcomes at 90 days. Larger confirmatory RCT is ongoing.

Broadening the time window for thrombolytic and other reperfusion therapy beyond 3 hours:

DIAS trial: Double-blind, randomized, multi-center, controlled trial that selected patients with diffusion < perfusion mismatch for randomization to one of 3 doses of the thrombolytic desmoteplase or placebo between 3-9 hours from ischemic stroke onset. The results demonstrated a dose-dependent trend on reperfusion 4-8 hours after treatment initiation and on 90 day clinical outcome. Statistical significance for both early reperfusion and 90-day outcome were demonstrated for the high dose group. DIAS is the first RTC to prospectively demonstrate (1) treatment vs placebo effects in a sample limited to penumbral neuroimaging pattern (2) clinical benefit of intravenous thrombolytic > 3 hours (3) dose selection based on MRI reperfusion outcome.The results for the companion DEDAS trial were consistent, although not statistically significant. DIAS-2 is an ongoing RTC of 2 the highest two doses vs placebo, selecting patients by penumbral pattern on MRI or CT.

DEFUSE trial: A multi-center, open label trial to investigate whether specific PWI/DWI profiles predict the clinical response to early reperfusion in patients treated with intravenous t-PA therapy between 3 and 6 hours after stroke onset. The results of this study demonstrated that baseline MRI findings can identify subgroups that are likely to benefit from reperfusion therapies and can also identify subgroups that are unlikely to benefit, or may be harmed.

EPITHET trial: Double-blind, randomized, multi-center, controlled trial to determine whether the extent of the ischemic penumbra apparent on perfusion-diffusion MRI can be used to identify patients who would respond positively and safely to tissue plasminogen activator (tPA) 3-6 hours post-stroke.

MR RESCUE trial: Initiation of a randomized, multi-center, blinded, controlled trial of large vessel occlusion up to 8 hours from onset to determine (1) if diffusion-perfusion MRI can identify patients who might benefit from mechanical embolectomy with a balloon catheter and retriever (MERCI device) and (2) if embolectomy is more clinically effective than standard medical management. Study enrollment is currently ongoing.

Carotid Occlusion Surgery Study (COSS): The failure of collateral circulation to maintain normal cerebral blood flow can lead to increased oxygen extraction. Increased oxygen extraction, as measured by PET, is a powerful and independent predictor of subsequent stroke in patients with atherosclerotic carotid occlusion. The COSS trial is a randomized trial of surgical bypass to improve cerebral blood flow versus best medical management in patients selected by PET measurements of increased oxygen extraction. Study enrollment is ongoing

New developments in stroke imaging diagnostic markers

Prospective blinded comparisons of MRI and CT have established the equivalence for acute ICH diagnosis and the superiority for acute ischemic stroke.

Early reperfusion with MRI, even partial reperfusion (e.g., at least 30% reduction in the volume of ischemia, the criterion in DIAS), with tPA and investigational thrombolytic therapies predict long term clinical recovery.

Increased recognition of the association of chronic cerebral microbleeds (CMB), seen on gradient echo MRI, with the risk and risk factors for ICH. The risk CMB of as a predictor of hemorrhagic complications of thrombolytic or antithrombotic therapy as well as the implications of CMB for risk factor modification needs to be determined.

Longitudinal studies of patients with silent strokes diagnosed by MRI have provided evidence accumulates that patients with silent brain infarcts and white matter lesions have increased risk of symptomatic stroke independent of other stroke risk factors and are more likely to develop global cognitive decline and dementia. Acute and subacute silent strokes diagnosed by DWI predict subsequent strokes, both symptomatic and asymptomatic, and DWI positive TIA indicate increased risk for subsequent stroke.

Imaging markers of risk of hemorrhagic transformation, especially in the setting of thrombolytic therapy, have suggested these markers as investigate approaches to reducing the hemorrhagic risk of thrombolysis. The depth of ADC reduction or early blood brain barrier seems with CSF enhancement on post gadolinium FLAIR MRI have each been clear associated with this risk. The association of blood levels of matrixmellatoproteinase-9 (MMP-9) with MRI features of infarct size, growth of DWI lesion volume, hemorrhagic transformation, as well as with thromobolytic treatment, opens this area to further correlations with peripheral blood markers.

Studies of stroke recovery combining functional MRI and transcranial magnetic stimulation (TMS) have provided a novel and promising approach to deducing mechanisms functional reorganization during stroke recovery.

Imaging of the potentially vulnerable atheromatous carotid plaques has advanced using a variety of imaging modalities. Ultrasound, CT, and MRI have respective advantages and disadvantages for imaging degree of stenosis, intima-media thickness, calcifications, fibrous cap thickness, and lipid core. An inflamed carotid plaque associated with recent clinical events may be imaged with 18F-fluorodeoxyglucose PET, and mononuclear inflammatory cells in experimental plaque take up MRI-detectable ultra-small superparamagnetic iron oxide particles (USPIO) injected into the circulation. Multimodal MRI has shown the most promise in differentiating the calcifications, fibrous cap, and lipid core of the plaque, which have been related to clinical events. The future of clinical in vivo plaque imaging promises molecular probes and targeted contrast agents to further refine the risk of a plaque beyond the percent stenosis and plaque morphology.

Major advances in non-invasive vascular imaging techniques (including advanced CTA and MRA techniques) now allow detailed imaging or the aortic arch, neck, and intracranial vessel atherosclerosis without the risks associated with conventional angiography. Noninvasive imaging techniques for aneurysm diagnosis and definition are also being applied to help determine the most appropriate choice of therapy as well as to aid surgical planning.

Penumbral markers and predictors of tissue fate.

This past 5 years saw a large assortment of models using MRI and, increasingly, CT approaches to predicting clinical outcomes and tissue infarction from acute anatomical, hemodynamic, and diffusion variables alone or in combination (see above). Virtually all published models have demonstrated moderately good predictive validity, but none has shown clear superiority for feasibility, accuracy, and validity. Standardization across centers and studies, unfortunately is lacking. Until definitive comparisons among predictive models are undertaken (see below) and one is validated on prospective, large, well-controlled multi-center samples as predictive of response to therapy, such predictive models-whether the simple qualitative perfusion>diffusion mismatch or sophisticated, automated multiparametric models-will be of limited utility for routine clinical decision making or studies of stroke pathobiology.

Mapping the penumbra and core with novel positron-emission tomography (PET) using the hypoxia marker 18F-labeled fluoromisonidazole (FMISO) has emerged as a promising tool.

General technical developments and clinical implementations

  • 64-slice CT scanners and 3 Tesla MRI scanners have become standard (high-end) clinical scanners, permitting greater speeds of acquisition and sensitivities for brain pathologies
  • CT angiography (CTA) and CT perfusion (CTP) of the brain have become clinically routine
  • MRI diffusion weighted (DWI) and bolus-tracking perfusion (PWI) of the brain, and contrast enhanced MR angiography (MRA) of extracranial arteries have become clinically routine
  • Multimodal CT and multimodal MRI are feasible as the initial emergency neuroimaging screens, and are routinely used as such at many centers.

Key Unresolved Scientific Questions from 2001 SPRG

Three neuroimaging priorities emerged from the first stroke PRG along with 17 specific areas for future research. The 3 main priorities identified in the first SPRG were:

1. Define the relevant pathophysiologic mechanisms in human cerebrovascular disease as determined by molecular and functional neuroimaging, and integrate this knowledge with appropriate data from experimental systems to develop new therapeutic approaches.

2. Identify and prospectively validate neuroimaging markers of tissue injury for prediction of clinical outcome in large patient samples.

3. Identify neuroimaging markers of potentially salvageable brain tissue in acute stroke.

As noted above, significant advances have been made in the second and third priorities, but priority 1 still remains largely unaddressed. In addition to this general priority, 3 or the 17 specific topics also stood out as still not sufficiently addressed. These 4 issues are listed below with suggested approaches to address these questions.

  • Define the relevant pathophysiologic mechanisms in human cerebrovascular disease as determined by molecular and functional neuroimaging, and integrate this knowledge with appropriate data from experimental systems to develop new therapeutic approaches.

    Development of molecular and cellular imaging in human cerebrovascular disease remains in its infancy but does offer promise as a technique to better understand stroke pathophysiology and in turn provide data to develop new therapeutic strategies.
  • Multicenter studies to determine if the application of diffusion MRI and magnetic resonance angiography, as opposed to CT, in acute stroke results in improved patient outcome and/or reduced healthcare costs.

    Although diagnostic superiority has been demonstrated for MRI, no formal studies have demonstrated that an MRI based approach affects patient outcome or healthcare costs. A multicenter study to address this question poses significant logistical and research challenges. A randomized trial would likely not be feasible or embraced by investigators. An alternative approach is to include matched hospitals designated as using either an MRI approach or a CT approach.
  • Studies to determine the neuroimaging correlates of post-hospitalization deterioration and mortality following intracerebral hemorrhage.

    This question could be addressed through multicenter, longitudinal, natural history studies of patients with primary intracerebral hemorrhage.
  • Studies of neuroprotective drugs, to determine if they can penetrate the ischemic penumbra in humans at sufficient concentrations to be effective.

    This question could be addressed by performing drug-labeled studies employing PET or MR Spectroscopy.

New Top 3 Priorities

Molecular and Cellular Neuroimaging: Develop new technologies to define pathophysiologic mechanisms


  • Image labeled cells or drug delivery (stem cell, inflammatory cells, neurogenesis, cell migration, therapeutic agents)
  • Imaging of atherosclerosis (detection of unstable plaque, carotid distensibility / wall thickness, carotid strain/stiffness, intimal-medial thickness)
  • Imaging of apoptosis, endothelial events

Identify and Validate Imaging Markers of Brain Injury and Prognosis


  • Clarify the accuracy and clinical utility of imaging markers of potentially salvageable brain tissue (markers of the ischemic penumbra: DWI/PWI, functional MRI, CTP)
  • Identification of subgroups most likely to benefit from acute stroke therapies and validate imaging markers as outcome predictors (which imaging modalities are optimal to achieve these goals)
  • Imaging of intracranial hemorrhage (natural history, role of edema, prognostic features, predicting SICH, predicting ICH recurrence risk, optimal imaging modalities)
  • Hemodynamic assessment: Autoregulatory function, oxygen extraction - for arterial occlusive disease, subarachnoid hemorrhage induced vasospasm
  • Imaging of global cerebral ischemia to predict functional outcome/recovery
  • Imaging markers of blood brain barrier injury
  • Refine noninvasive techniques for aneurysm diagnosis, definition and treatment

Functional imaging of patients to clarify mechanisms of recovery of function and enhance recovery


  • Visualize and understand basic mechanisms of brain recovery (functional image guided transcranial magnetic stimulation to promote recovery)
  • Develop tools to image axonal growth and synaptogenesis
  • Early prediction of outcomes or identification of subgroups most likely to have therapeutic benefit
  • Role of multimodal imaging (PET/CT, MR/PET)

New research areas to watch:

  • Four dimensional MR imaging
  • Arterial Spin Labeling assessment of brain perfusion
  • MR methods for assessment of tissue oxygenation
  • Optical imaging for functional activation, ischemia
  • Infrared spectroscopy for real time assessment of cerebral perfusion
  • MEG/fEEG
  • Imaging small intracranial vessels
  • Imaging of cortical spreading depression
  • Prediction of cardiac dysfunction following acute brain injury
  • Risk factors for aneurysm rupture (configuration, growth, pulsation patterns)

Resources Needed and Approaches to Address these Questions

  • Financial support to create infrastructure for stroke neuroimaging research including dedicated scanners and cerebrovascular neuroimaging research centers.
  • Multicenter collaborations to conduct prospective hypothesis-driven studies as well as retrospective analyses of pooled datasets of all types of cerebrovascular disease (including, ischemia, intracerebral hemorrhage, atherosclerosis) to 1) better define the pathophysiology of stroke and recovery; 2) study the natural history of stroke and recovery; 3) identify and refine optimal surrogate outcome markers; 4) develop predictive models that can be used to assist in treatment decisions; and 5) conduct phase II clinical trials employing imaging for patient selection and as a surrogate marker to identify promising therapies.
  • Standardization of image analysis terminology and methodology (e.g. MR perfusion imaging) including support for NIH workshops to allow neuroimaging leaders to define optimal approaches to image processing and analysis.
  • Development of a centralized, electronic neuroimaging repository to allow analysis of large, pooled datasets.
  • Development of formal training programs (e.g. fellowship) for cerebrovascular imaging research.



Co-chairs: Patricia Hurn PhD, Ray Swanson

Members: Kyra Becker, Pak Chan, Sue Duckles, Marc Fisher, Giora Feuerstein, Mark Goldberg, Jialing Liu, Richard Traystman

NINDS Liason: Dr. Eugene Golanov, Michael Moskowitz

Process: The following is a distillation of serial email polling and conference call amongst committee members to address the 4 key areas for SPRG2 review. Members identified items to be included in each key area list (approximately 112 resulted), then were asked to champion their choices in the conference call. The next step was a group vote to rank all candidate items as most important for attention of NINDS. In the document below, only these selected items are displayed and generally ordered so that items listed first in the category held heaviest weight and agreement within the group deliberations. In the original SPRG, the working group for Neurovascular Protection focused very strongly on getting a new concept in play, i.e. the neurovascular unit. As discussed below, this has done well in the NINDS community and stroke community at large. For SPRG II, discussion was not restricted to this concept although its validity is not questioned. A key area of focus was on translational approaches to stroke research and associated progress.

Key Areas for Review:

1. Seminal advances in this area since 2001:

  • Successful change in conceptual framework for stroke pathophysiology, i.e. that the unit of analysis is a neurovascular unit composed of linked cells of neuronal, glial, vascular origins that are functionally connected in matrix fashion. This concept has enabled the community to move away from purely neuronal models of neuroprotection or endothelial models of vasoprotection, without sacrificing molecular hypotheses. Once indication of the success of the concept is the number of publications about protection or plasticity of the neurovascular unit in high profile journals. Equally common at present is a wealth of presentations relevant to the neurovascular unit within national and international neuroscience meetings.
  • Acknowledgement of the important role of inflammation in stroke pathophysiology and that inflammatory responses are viable therapeutic targets for stroke prevention, treatment and recovery
  • Creation of the Stroke Academic Industry Roundtable (STAIR) recommendations and wide use of STAIR recommendations in pre-clinical evaluation of neuroprotective and recovery enhancing drugs
  • Recognition and intense investigation of matrix metalloproteinases (MMPs) and their role in blood-brain barrier function and dysfunction, on interaction with tPA (e.g. MMP-9), and role in post-injury neurovascular remodeling.
  • Encouraging success of the free-radical scavenger NXY-059 in clinical trial for neuroprotection. The importance of this positive clinical trial is not only its effectiveness for stroke therapy but its emphasis on the importance of pro-oxidant mechanisms in the pathophysiology of the disease over time.
  • Continued growth in our understanding of growth factors, not as neuroprotectants, but as factors that enhance mechanisms of plasticity. This area has been particularly important in the newer fields of neurogenesis and angiogenesis, where growth factors are vital in supporting these processes. Key members of growth factor families, such as VEGF and G-CSF, provide neuroprotection and enhance neurogenesis. As discussed below, combinational therapeutic mechanisms are believed to be of great value in treatment of stroke.

2. Unresolved questions/issues from SPRYG1

In general, there has been progress toward many of the previous recommendations and priorities from the original SPRG, however a number of areas relevant to neurovascular protection remain under-addressed. We reviewed the April 2002 and March 2004 SPRG reports of high priority programmatic activities. Scientific priority areas 2 and 3 are exemplary in how limited progress has been to date. Priority 2 is to study the interface between vasculature, and cellular, matrix, and hemostatic mechanisms, to achieve understanding of hemorrhage and infarction. Priority 3 is to understand blood flow and perfusion optimization. In response to these two priorities, four actions emerged. One was a training workshop for treatment of intracerebral hemorrhage. The other three appear in the form of program announcements (PAS-03-165; PAS-04-072) and RFA-HL-05-004. From these three implementation efforts, a total of 6, 11 and 5 stroke related grants were awarded, respectively. In addition, we reviewed the resources priorities in SPRG report, and several were detailed that are relevant to neurovascular protection. Number two priority is toward the development of large and small animals models that reflect the complexity and diversity of the human brain and its response to stroke. In response to PAR-01-077, NOT-NS-02-013, NOT-NS-04-009, not a single stroke grant was funded. Number three resource priority is to expand brain imaging capabilities and again implementation initiatives resulted in no funding of stroke related grants. A single stroke related grant emerged in response to PA-02-003, PA-04-006 and PA-06-279. The reason for this apparently small number of awards relevant to priority and resources areas as listed in the original SPRG is multi-factorial. However, we believe that a likely contributing factor is at the level of the study section. Although a variety of funding mechanisms appear to have been offered, there appears to be little effect as the applications go through the peer review process. This problem may involve assignment of grant applications to study sections not specifically poised to evaluate responses to SPRG initiatives or that composition of existing study sections include members who are not excellent ambassadors for the stroke field. For example, if a reviewer is focused on a precise sub-molecular target, his/her interest will be limited on encountering applications intended to develop stroke animal models that are heterogeneous in their genetic background, involve age or specific compounding diseases.

  • A central issue arising from SPRG that remains unresolved is that neurovascular protection studies rarely are carried out in aging brain animal models, which is the primary arena for stroke. In addition, few studies use models that account for sex-specific effects in mechanism or outcome. Both factors, age and sex, could readily impact translation of pre-clinical data to human trials of women and men.
  • Lack of attention to signaling pathways involved in maintaining or restoring the integrity of the blood brain barrier. This has been an area where the confluence of mechanisms involving the endothelium, astrocytes and microglia in the neurovascular unit remains understudied.
  • More emphasis is needed on evaluating long term endpoints (e.g. 2 months or longer in pre-clinical studies), particularly functional endpoints to appropriately assess motor/ behavioral paradigms or imaging endpoints with the resolution to evaluate the microcirculation (i.e. 10 m vessels)
  • Inadequate utilization of models that evaluate stroke on the background of complex disease or risk factors (e.g. obesity, diabetes etc) or are designed to evaluate invasive devices as opposed to pharmacological interventions
  • Lack of appropriate vehicles to deliver protective materials across the BBB (other than traditional small chemical molecules)
  • Inadequate focus on white matter injury at all levels, risk factors, pathophysiology, and treatment
  • Failure to develop "glioprotection", i.e. the contribution of glial cells to the neurovascular unit and in mechanisms of repair

3. Top Priority actions needed (by the NIH)

  • As discussed in key area 2, there must be a novel approach to peer review of proposals launched by SPRG initiatives and priorities. This may be the designation of a standing study section to evaluate proposals or incorporation of a cadre of reviewers who serve on study section when such proposals are present (i.e. the concept of ombudsmen for SPRG).
  • Support for development of animal models that include key factors of age, sex, functional outcomes, underlying disease processes that accompany stroke in humans, interventional device trials
  • Support for projects that highlight use of collaborative research with investigators outside our field and provide incentives for multidisciplinary research combinations. This is particularly important for activities that promote translation of basic research to clinical treatments.

In addition to these top three priorities, we would highlight the need to create and support cross disciplinary training and interaction between clinicians, basic scientists, and clinical trialists. This should include funding of semi-annual/annual meetings of multidisciplinary working groups that are tasked to serve over 2-3 years and investigate impediments to the field.

4. New research areas / opportunities

  • "Combotemporal stroke therapy". It is well recognized in the community that effective interventions likely require more that one drug to be delivered either simultaneously or in sequence. Effective treatments likely involve multiple agents that have multi-function to target neurovascular protection, preservation and plasticity. Such areas of research are essential but are not well received in traditional study sections.
  • Accelerate studies of neurogenesis and angiogenesis in recovery after stroke, particularly in combination with stem cell therapy
  • Focus on the immune response (CNS and peripheral) and its consequences following blood brain barrier breakdown and of post-stroke immunodepression
  • Pursue further aspects of successful NXY-059, particularly if there are connections to the use of tPA and to ameliorating undesirable hemorrhagic complications of tPA
  • Elucidate sex differences in pathophysiology of ischemic brain injury
  • Define role of metabolic syndrome in neurovascular responses to ischemia
  • Using the injured blood brain barrier and its altered transporters as a delivery systems for large molecule, neuroprotective agents or cells



Co-Chairs: Valina L. Dawson, Jonathan Rosand

Members: Alison E. Baird, Frank C. Barone, Ulrich Dirnagl, Robert E. Gerszten, Laura Lubbers, David J Stone

NINDS Liaison: Katrina Gwinn-Hardy, Thomas P. Jacobs, Melinda Kelly

Statement Of The Problem - Progress To Date

Application of discoveries made using platform technologies have the capacity to lead to new approaches in the management of patients. The sequencing of the human genome, the advent of microarray technology, advances in proteomics (analysis of protein expression) and metabolomics (analysis of metabolites) and in particular, advances in bioinformatics are opening up new horizons for studying stroke in the both the experimental and clinical setting.

Currently platform technologies are used largely in experimental settings and some technologies are still in development. Recent technologic advances are creating new opportunities and it is likely that platform technologies will increasingly be applied beyond the experimental study of stroke and move into stroke diagnosis and treatment.

The complexity of cerebral vascular disease and the complex processes involved in stroke recovery, pose challenges to the application of platform technologies. Optimal exploitation of new platform technologies will require both large numbers of highly phenotyped patient samples along with a diverse range of control samples. New targets and pathways must be validated in the context of both animal models and human systems. To this end a better vertical validations of animal models as predictors of human outcomes are needed. In order to meet the challenge of advancing patient care through the use of platform technologies, a strong focus on experimental design, statistical, and bioinformatic analysis along with free access to data sets is essential.

Although results are preliminary in nature the immediate potential of the "omics" methods in the clinical setting appears to be in the provision of "biomarker" information with unique array signatures detectable for acute ischemic and hemorrhagic stroke. The "omics" methods may also permit the identification of new biological targets in the human that have not been identified or are not able to be identified in experimental stroke models.

Seminal Scientific Advances since SPRG 01:

  • Wide acceptance of standards for presentation and analysis of Genomic data
  • Increased accessibility of platform technologies
  • Advances in bioinformatics
  • Completion of the first gene expression, proteomic and metabolomic studies in patient-derived samples

Key Unresolved Scientific Questions:

  • Lack of widely accepted standards for collection, presentation and analysis of proteomic and metabolomics data
  • Recent development and acceptance of the "MIAME format" for genomic data is an encouraging advance, but similar initiatives are needed for proteomics and metabolomics studies.
  • Limited overlap of gene expression signatures between patients and animal models
  • Identification of molecular markers of stroke and stroke risk


Genomic technologies have rapidly advanced and are widely used. Genomics is the study of gene expression in different physiologic and pathophysiologic states. Over the last 2-5 years genomics has been applied to the study of stroke in experimental settings to explore the response of brain tissue to ischemia and reperfusion injury and towards the investigation of the phenomena of preconditioning (i.e. endogenous protection) against neural injury. In experimental stroke models gene expression signatures of acute stroke have been identified in the peripheral blood. Subsequently, the first reports of gene expression profiling of the peripheral blood in the clinical setting of ischemic stroke have been reported, showing a gene expression signature of ischemic stroke that is moderately reproducible. Comparison with other disease states suggests at least some specificity of blood-derived gene expression changes for ischemic stroke in humans. It has been noted that there is only a small overlap of blood-based gene expression signatures between man and rodent. There also seems to be a partial dependence of the human ischemic stroke and genomic blood "stroke-signature" on the presence and extent of associated vascular risk conditions.

Functional genomics, as the name implies, focuses on genome-wide analysis of gene function and/or effect on phenotype. Tools such as siRNA libraries can be used to screen literally thousands of genes for effects on in vitro phenotype in a high throughput assay. Therefore, the expression to function connection does not need to be deciphered by the scientist, as the assay readout itself is the desired function/phenotype. These sorts of screens are currently only limited by the phenotypes (e.g. transcript changes, secreted proteins, etc.) that can be modeled in a readily transfectable cell line, but with the development of lentiviral shRNA libraries, the potential for multiple cell-type and in vivo screening are being realized. These libraries will most likely not be as useful as microarrays or proteomics for the detection of biomarkers, which need to correlate in expression with a given phenotype, as opposed to inducing the phenotype. Thus by understanding which platform (functional or expression) is appropriate for answering specific questions, researchers will be better able to take full advantage of the developing genomic technologies.


While genomics-based technologies are useful tools and have created advances in understanding of disease processes the ultimate goal for many researchers is the identification of proteins that play a key role in prevention, progression of, or demarcation of disease. Thus, there is great interest in developing methods to broadly profile the proteome in basic research, drug discovery and clinical settings.

Current methods of proteomic profiling can be divided into two different categories: (1) targeted profiling of single or large panels of known proteins and (2) global proteomics, which aims to identify all protein or peptide sequences that change in response to a stimulus. Researchers have long targeted single, known proteins of interest using various immunoassay approaches. This has the advantage of great sensitivity (ng/ml to even fg/ml range) but the disadvantage of assessment of one analyte at a time. Newer protein chip approaches, have expanded the number of analytes that be measured within the same sample, saving countless hours at the bench while retaining immunoassay sensitivity. This approach is being applied to clinical biomarker programs in other Neuroscience-based disease areas, but no reports exist to indicate it is being utilized in the study of stroke.

Global proteomics profiling for novel protein discovery is an active area of interest and debate. Different approaches to protein separation have been used including 2D gel electrophoresis but more recently samples are processed without gel separation using microcapillary high-performance liquid chromatography (HPLC) systems coupled to mass spectrometers. Tandem MS/MS and affiliated software can then used for differential quantification and molecular identification of peptides using publicly available databases found at sites such as and

Pilot proteomic data in acute ischemic stroke have also been reported. Currently a panel of 4 proteins is under investigation as a potential diagnostic test for acute ischemic stroke. Pilot data using the surface-enhanced laser desorption/ionization (SELDI) methodology has also identified proteins associated with ischemic and hemorrhagic stroke. The first blood based biomarker for predicting stroke risk associated with atherosclerosis was approved by the United States Food and Drug Administration - lipoprotein-associated phospholipase A2 (this plasma based biomarker was also approved for the prediction of future risk of myocardial infarction).


Metabolomics is the study of metabolic changes and/or metabolites. The identification of biochemical signatures is actively being investigated as a biomarker method to facilitate the triage and treatment of patients. Methods have been used to identify novel biomarkers of myocardial ischemia in a group of subjects undergoing cardiac stress testing, half of whom demonstrated inducible ischemia. Metabolomics methods apparently have not yet been applied to the problem of stroke.


In addition to the need to understand the statistical significance of "omics" data, is it is critical to put the information gained into biological context. This can be a daunting task for researchers because of the very large datasets that are generated from investigations using platform technologies. Bioinformatics provides computational tools for data basing large data sets and the tools access and analyze these data sets regarding, gene expression, gene products, protein expression and post-translational events, pathways and biological networks. There are several skill sets have evolved in the Bioinformatics Science including the use of sequence/structure processing approaches such as gene string algorithms, data management and functional annotations. As larger more complex platform technologies have emerged the field of bioinformatics has evolved new approaches to understand and characterize the data sets including learning algorithms, ontologies, dimensionality, applying advanced statistical approaches including multi-variate and principal component analyses, and new data visualization tools. With the need to develop systems/network approaches to understand complex biologies, additional tools have evolved including data integration, mining tools and use of pathways/network theory. These pathways/networks can be "canonical" or "inferred" involving novel pathways or these pathways can be apparent working networks derived from platform data, the literature and available data bases that can be difficult or challenging. As the platform technologies continue to develop the technology will become available to more investigators advancing the need for additional bioinformatics tools that are powerful but also accessible to investigators outside the bioinformatics field. Currently the development of curated pathways analysis tools (e.g., Ingenuity, now aid significantly in modeling of the genomic data, for those with the resources to access these tools. These tools identify possible relationships between regulated genes and derive gene networks that can be the basis of subsequent hypothesis testing. Development of tools that are widely available to academic and private researchers would aid in analysis of what can be seemingly long lists of unassociated genes or signaling events. Nonetheless, while gene networks can provide insight into possible mechanisms of, or response to disease processes, they do not replace functional testing and hypothesis driven work.

Current issues are:

Challenges And Questions

New key research areas:

  • How best to select and characterize the patients and animal models studied
  • How to manage, evaluate, standardize the enormous amounts of data created and combined in platform studies
  • Need for multidisciplinary teams of biologists, stroke clinician-investigators, bioinformaticians, and statisticians dedicated to defining the questions that can be asked using these technologies.


As genomic technologies become more readily available, it will be critical for researchers to understand the proper application and the limitations involved. For example, researchers looking for new drug therapy targets in stroke may pursue them via microarray studies. While it is possible to uncover drug targets this way, it is not clear that targets for a given condition are more or less likely than other random and reactive genes to be differentially regulated. Similarly, insults such as stroke or animal models of stroke generally result in large gene expression signature, sometimes consisting of hundreds to thousands of genes. This leaves researchers with a large list for which it is impossible to do an exhaustive follow up, and for which most likely 1% or less are viable targets. New understanding of pathways involved can be achieved through a systems biology analysis, though this can be time-consuming and requires considerable computing power. The problem for the average researcher in this case would be that an expression genomics platform was used to answer a functional genomics question.

Peripheral white blood cells are, at least in part, a surrogate tissue of ischemic stroke, and so it is important to determine how specific are changes in gene expression in peripheral blood for cerebral ischemia. Also, it is possible that gene expression changes may in part be a reflection of generalized vascular disease, as atherosclerosis is a chronic inflammatory disorder. Therefore, the gene expression changes of acute stroke need to be tested against acute stress conditions, against other vascular diseases, and viewed in relation to stroke mimics such as seizures, brain tumors and migraine, and perhaps psychiatric diseases. The specificity of novel protein markers for acute ischemic and hemorrhagic stroke also needs to be determined. Additionally there is a technical need to increase the sensitivity of the assays so that smaller sample volumes can be assayed to decrease cost so that a larger patient population can be assessed. A decrease in cost is also critical for moving this technology to the clinic for triage of patients.

A number of practical aspects relate to genomic studies in the clinical setting. These include: (1) the volume of blood draw (2) the type of blood tube(s) into which the samples should be drawn (an example, for gene expression profiling, the PAXGene method has low sensitivity but permits stabilization of RNA for up to 24 hours) (3) the availability of specialized clinical laboratories for isolation of RNA from blood including peripheral blood mononuclear cells, (4) storing mRNA for real time PCR studies, (5) the cost of microarray chips and (6) the need for bioinformatics expertise. This last conern is probably the most limiting aspect of any study and is beyond the scope of many investigators. For studies of this complexity it is critical to have expert collaborators across multiple medical and scientific disciplines.


A limitation that continues to plague global proteomics profiling is the relative lack of sensitivity and difficulties in analyzing complex samples such as blood and CSF. However, immunodepletion of abundant proteins such as albumin & IgGs from the sample can improve sensitivity as these entities mask less abundant proteins, which are often of greater interest. Sensitivity also can be improved by various sample fractionation approaches including: molecular weight fractionation of samples, isolation of sub-proteomes such as the N-linked glycoproteins or cysteine-rich peptides. Currently, technologies of this complexity are offered by commercial operations however, key drivers of technological advancement continue to be within academia.

Standardization of data bases and data presentation is needed. The human proteomics project ( is making strides in standardizing proteomic approaches across laboratory settings. Specifically, the HUPO Proteomics Standards Initiative (PSI) defines community standards for data representation in proteomics to facilitate data comparison, exchange and verification. The PSI also has outlined standards for key areas of proteomics including mass spectrometry data, and has developed a general proteomics format for the full representation of a proteomics experiment.

Proteomics is being used successfully in other research field to identify biomarkers of disease and its progression. To fully realize the utility of proteomics there is a need for appropriate training in methodology, statistical methods for handling large data sets, and means to put the data into context using pathways analysis tools. As yet, there are no published reports on the use of global proteomics profiling in stroke. Currently, the availability of the technology required for proteomic profiling appears to be limited and as such these approaches are not yet widely used by the larger research community. This situation can be expected to change as resources are acquired by research hospitals and academic groups. Identification of proteins and pathways that change prior to, during, or in the recovery phase of stroke will provide new insights into the disease process and new opportunities for therapeutics. Successful application of this approach will require standardization of sample collection and preparation (i.e. generic to all clinical studies) and inclusion of appropriate control (e.g. stroke mimics, high-risk non-stroke) populations for comparison. Most importantly, persons with appropriate training in execution of the techniques, experimental design and interpretation of the data will be cornerstone of progress made with this approach.


Metabolomics methods have not yet been applied to the problem of stroke (to the best of our knowledge). Sampling is a problem as the tissue immediately affected cannot be sampled readily but screening studies of peripheral blood may reveal signatures that could be developed for rapid analysis of patient status and thus facilitate patient care as this platform technology develops.


As genomic tools become more readily available to basic and clinical research labs, a deeper understanding (and thus training) will be necessary in certain areas if scientists are to fully and correctly utilize the tools. Currently most studies using genomic strategies such as microarrays or functional genomics such as siRNA libraries are collaborations between disease area experts (stroke researchers and clinicians) and genomics experts (in core facilities, CROs or genomics departments). In this situation, the disease-area expert does not need to be an expert in advanced statistical methodologies; these concerns are addressed by the genomics staff, and issues with false positives, artifacts, etc are addressed and avoided to the extent possible. However, it has been the case with almost all technologies known to date that with time, cost is reduced and availability increases; there is no reason to believe that this will not also be the case with genomic technologies. One can easily envision a time when microarrays, siRNA libraries, and proteomics are as commonplace as Western blots. Most scientists, however, have not had training in the methodologies necessary for the analysis of large datasets. The methodologies are not beyond the intellectual reach of most MDs and Ph.D.s; rather it is that if the tools are not correctly understood, misapplication can result in the incorrect interpretation of data, in turn resulting in both time and money being lost in the pursuit of false leads.

An excellent example is the use of advanced classification methods such as stochastic discrimination and random forest. These methodologies use large numbers of data points (for example, mRNA levels as measured on microarrays) to classify patients/subjects into discrete groups (stroke vs. TIA vs. control). Unlike some methods, which use a small number of "markers" to establish a classification, these use a large number which preferably are not correlated with each other (as less information is gained by the addition of a component which is correlated with others used in the model). Before the model is built, the dataset is usually filtered on some basic statistical measure (such as a t-test for significance between the groups to be classified, in this case stroke vs. control). Once a model has been built, it is re-tested or confirmed on a separate population, to ensure that the data has not been "over fitted" and that the classifier developed is truly useful. A common practice when the number of patients is limiting is to validate the model by a "leave-one-out" analysis. In this case, the random forest classifier is developed multiple times, with one subject removed (thus the number of iterations will equal the number of patients, i.e. run once for each patient with that patient missing). Each time the classifier is built, it is then tested on the remaining/removed patient to determine if he/she is classified correctly. This type of validation is attractive as it reduces the number of individuals need to get an estimation of the selectivity/specificity of the generated model. This method, however, is frequently misapplied, and multiple cases can be found in the literature. A common mistake is to filter the dataset based upon all individuals, and run the leave-one-out analysis on the remaining dataset. This results in an over-estimation of the accuracy of the developed model, as it is not a true leave-one-out. The individual must be left out of the entire run (and therefore be left out of the filtering step), otherwise the result is only measuring the degree of filtering, not the true variation in the population.

This example shows how a misunderstanding of how the statistical model works can result in a misapplication and an erroneous result. Note that it is not important to fully understand the mathematics behind the model, but rather to understand the principles behind it and its applications. Most scientists cannot derive the formula for a 2-way ANOVA from scratch, although most can interpret one's results and determine if it has been applied correctly. This level of understanding will be necessary for anyone attempting to run genomic/genome-wide analysis in their lab without the help of a core facility or collaborating statistician.

A number of the necessary tools have become readily available to scientists; for example various clustering algorithms (agglomerative, K-means, etc) which are extremely useful in the analysis of microarray data, are becoming widespread through commercially available data analysis packages. There are various other techniques such as linear discriminant analysis and principle component analysis which are useful for identifying trends in large datasets; making basic training in the application of these methods available will greatly enhance the ability of individual researchers to fully exploit the power of genome-wide analysis. Similarly, there are aspects of common statistical procedures that can be applied differently, such as the estimation of a q-value (false discovery rate) as compared to a p-value (false positive rate). Both are useful measures, and it lies with the researcher to know when each is appropriate in large datasets.

Optimizing patient selection for clinical studies:

Stroke is the acute manifestation of a chronic progressive disease. When, patient material is sampled, during the course of disease, restricts the scientific questions that may be asked. For example, proteomic profiles of samples collected at the time of acute stroke are likely to be altered as a result of the acute event and may offer little insight into stroke susceptibility. In contrast, samples obtained prior to the development of stroke are likely to facilitate scientific questions regarding stroke susceptibility.

In human and animal studies, choice of the phenotypes for investigation is likely to affect the success of the investigation. Just as genetic investigations are most likely to be fruitful when the phenotypes being studied are known to be heritable, there may be patient characteristics more likely to correlate with genomic, proteomic, or metabolomics signatures. One example of this in cardiac disease is inducible ischemia on exercise testing.

Research And Scientific Priorities

The problem of sampling and the specificity of blood derived signatures for ischemic stroke and for assessment of stroke and vascular disease risk. Because the affected tissue cannot be serially sampled before, during and following a vascular accident it is necessary to sample peripheral sources such as blood or CSF for clues to the events occurring in the ischemic neural tissue. Gene expression and proteomic (and metabolomic) changes occurring during acute stroke need to be compared against other acute stress conditions, along with other vascular diseases, intracerebral hemorrhage, and other neurological disorders and stroke mimics such as seizures, brain tumors and migraine. One approach could be in a large scale study involving a large group of patients with multiple medical disorders (problem is the cost). If rapid assays can be developed then a cost-effective approach would be to test out these assays in the field. Another approach is to use a controlled setting in which "omics" patterns are studied pre and post intervention: the situations that most closely resemble a "stress test for the brain" are cardiac bypass surgery and carotid artery surgery (or stenting) but these have the limitations of complexity and potential for biases from the intervention itself and also the relatively small number of events associated with each intervention. The refinement of functional neuroimaging of cerebrovascular disease may offer an appropriate phenotype (i.e., a form of stress testing) for study in stroke patients. An additional consideration for phenotype selection may involve the preferential selection of quantitative phenotypes over simple dichotomous classification of patient samples. Testing in a large animal model of stroke could be another approach. In terms of studying stroke and vascular disease risk, the addition of serial "omic" sample collection to ongoing large scale epidemiological studies could be considered (question is how long RNA remains viable for at -80 degrees) - this is also mentioned in section 4. Another more short-term and cost-effective approach is to conduct a study using an intervention that has been shown to reduce vascular risk.

Do study methodologies need to be standardized?

It should be noted that some of the practical issues referred to above (i.e. in relation to gene expression profiling) are being addressed by other groups. An example is "The Inflammation and Host Response to Injury, Large-Scale Collaborative Research Program" (so we should not re-invent the wheel). It is noted that the development of simple and rapid assays could bypass some of these problems. Consideration should be given to putting a working group together.

Evaluation of the usefulness and timeliness for the use of platform technologies in clinical or epidemiological studies?

There is significant potential for "pharmacogenomics" applications, but these are yet to be proven. Should "omics" sampling be incorporated into clinical trials (e.g., identification of predictors of response to treatment)? Should "omics" sampling be incorporated into epidemiological studies (e.g., the Framingham study or the Cardiovascular Health Study) and what should be the time frame and frequency of measurements?

Establishment of central databases and consortiums:

Who should be doing this work? Should consortiums be established and/or supported? How do young and or new investigators access the required technology and training?. There is an immediate need for bioinformatics expertise which is difficult to obtain. There is also a need for collaborations between scientists and clinicians, so that the most clinically useful information is obtained from these studies. Examples of stroke-associated network/pathway analyses, and new information on cellular signaling and metabolic changes in stroke and stroke risk are required. Such new information, if made accessible, can eventually contribute to new discoveries and the development of new assays, if we focus on newly discovered useful and reliable signatures/information.

Investigation and validation of vertical experimental modeling to human stroke patients?

The power of experimental modeling is only realized if the models predict outcome and response of human disease. Since stroke is a consequence of loss of blood flow to the brain it seemed straightforward to design animal models of vessel occlusion such as the MCAO rat and mouse models. These models are reproducible and have been used to test and validate cell signaling events. These models have not been predictive of the human response to therapy. In part recent reports address the issue that most experimental stroke studies are not designed to test efficacy. There is a need for further development of animal models that are closer to the human situation such as animals with chronic atherosclerosis, hypertensive animals, or animals with white matter disease to be used as additional screens in the preclinical setting. Additionally, preclincial experimental stroke studies need to be designed and powered to determine efficacy.


While it is possible to bring together stroke clinician-investigators with geneticists, systems biologists and bioinformaticians to exchange ideas, substantial advances in the application of novel genomic, proteomic, metabolomic and bioinformatic approaches to cerebrovascular disease will require sustained collaboration. Such collaborations, in which stroke investigators and scientists "learn each other's languages," can be facilitated by development of physically distinct thematic centers on academic medical centers where stroke clinician-investigators, scientific investigators, and clinician-investigators from other fields such as cardiology are housed together.

To prepare their graduates for this kind of collaborative scientific work, vascular neurology training programs must provide incentives and protected time for training not only in novel research techniques, but, more importantly, in how to best take advantage of these tools for the study of cerebrovascular disease. Instruction in the clinical application of these tools is likely to come from investigators in diseases other than stroke as well as from the scientists applying them from a fundamental perspective. As part of the phase of career development that follows clinical fellowship, Departments of Neurology must therefore support junior faculty trained in this collaborative approach, facilitating their placement in non-departmental space, while creating a departmental culture within neurology that celebrates and welcomes such faculty who will, no doubt, be required to spend substantial time away from the department itself.

Fostering long-term collaborations will also require recognizing and crediting on equal levels the value of diverse contributions to the research enterprise from all collaborators. Such contributions can include carefully characterizing patient samples, technical expertise and research ideas. Such recognition could come in the form of sharing the title of "principal investigator" on collaborative grant applications as well as the specific recognition of collaborative contributions by academic promotion committees. By supporting the development of long-lasting wide-ranging cross-disciplinary collaboration through the career of stroke clinician-investigators, funding agencies and academic institutions will ensure that the field remains continually open to novel technological advances.

As genomic tools become more readily available to basic and clinical research labs, a deeper understanding (and thus training) will be necessary in certain areas if scientists are to fully and correctly utilize the tools. However, it has been the case with almost all technologies known to date that with time, cost is reduced and availability increases; there is no reason to believe that this will not also be the case with genomic technologies. One can easily envision a time when microarrays, siRNA libraries, and proteomics are as commonplace as Western blots. Most scientists, however, have not had training in the methodologies necessary for the analysis of large datasets. The methodologies are not beyond the intellectual reach of most MDs and Ph.D.s; rather it is that if the tools are not correctly understood, misapplication can result in the incorrect interpretation of data, in turn resulting in both time and money being lost in the pursuit of false leads.

Support of short courses and integrated meetings, perhaps to occur as part of established national meetings, will facilitate the transfer of knowledge and training in these sophisticated technologies and applications. Such meetings could bring together fundamental researchers in the technologies, clinician-investigators from fields outside of cerebrovascular disease with track records of bringing platform technologies to the bedside, and clinician-investigators with expertise in cerebrovascular disease. The goal would be to alert cerebrovascular investigators to the potential applications of platform technologies, the proper design of experiments and analysis of data, while offering guidance on how to build collaborative teams in this area.

What Needs to be Done?

  • Harmonization criteria for phenotyping and data collection, presentation and storage for platform studies
  • Training programs, including short courses at meetings, that bring together interdisciplinary audiences of biologists, bioinformaticians, clinician-investigators and trainees to address application of platform technologies to cerebrovascular disease
  • Encourage (require?) consensus on data entry, format for NINDS-funded platform
  • Richly phenotyped and characterized patient samples with attention to quantitative traits that will yield maximum information for exploratory analyses.



Co-Chairs: S. Claiborne Johnston, Dilip Pandey, Bruce Ovbiagele

Members: Cheryl Anderson, Oscar Benavente, Don Easton, Mitch Elkind, Edward Feldmann, Heather Fullerton, Scott Kasner, Laura Sauerbeck

NINDS Liaison: Richard Benson, Scott Janis

Scientific Advances

  • Formation of JCAHO and state health department certified stroke centers, which have focused attention on quality improvement in secondary stroke prevention. More centers are systematically collecting data on prevention treatments after stroke and TIA, and these data are being used to study predictors of adherence and to develop novel interventions to improve implementation.
  • Paul Coverdell National Acute Stroke Registries have advanced understanding of barriers to secondary prevention strategies. They have also shown that use of standardized orders in stroke patients is associated with greater adherence to proven secondary prevention practices, and have identified predictors of failed adherence (such as failure of early initiation) that may be targets for future intervention studies.
  • Several studies in diverse populations have confirmed that the short-term risk of stroke after TIA is high. Clinical and imaging predictors have been identified and validated in independent cohorts.
  • The overall risk and risk factors for subarachnoid hemorrhage in patients with intracranial unruptured aneurysms have been identified (ISUIA). These risks have been compared to treatment risks with surgical clipping and coil embolization.
  • Several studies have identified imaging findings associated with subclinical cerebrovascular disease, such as white matter disease and silent infarcts, and have shown that they are associated with risk of clinical stroke (CHS, ARIC, Rotterdam, and Framingham studies)
  • Several studies have identified vascular risk factors as predictors of cognitive decline and dementia (CHS, ARIC, Rotterdam, and Framingham studies)
  • The high risk of stroke in patients with symptomatic intracranial atherosclerosis has been demonstrated and several risk factors (such as more recent symptoms and higher grade stenosis) have been identified.
  • NINDS-sponsored harmonization conference (04/2005) developed data elements to be used in characterization of patients for the study of vascular cognitive impairment, thus providing a basis for future research and comparison across studies.
  • Children with arterial ischemic strokes have been found to be at high risk for recurrence and several predictors of risk have been identified (such as vascular abnormalities).
  • Several large cohort and family studies have delineated the genetic and environmental contributions to risk of stroke subtypes and to white matter disease.
  • Publications from DECODE genetics in Iceland identified two novel candidate genes for ischemic stroke, and the manipulation of their gene products are being tested as possible therapeutic strategies.
  • New translational research has identified novel target for prevention (e.g., inflammation, metabolic syndrome) and new potential approaches (e.g., vaccination)
  • Major prevention trials have been published including several funded by the NINDS
  • Other sponsors: ACST, ACTIVE-W, ALLHAT, ASCOT-LLA, CHARISMA, ESPRIT, HPS, ISAT, LIFE, MATCH, MOSES, SAPPHIRE, SPARCL, SPORTIF, Women's Health Study (Primary Prevention with Aspirin), hydroxyurea for secondary stroke prevention in sickle cell children who have received long term transfusion (pilot trial; Ware et al)
  • The International Pediatric Stroke Study has been formed and has demonstrated the potential to bring together a large number of sites as the basis for recruitment to future clinical trials and the source of information on risk stratification.
  • A mechanism for warfarin resistance has been discovered with the demonstration of the key role played by vitamin K epoxide reductase complex subunit 1 (VKORC1)
  • The NINDS CRC is engaging practicing neurologists to encourage participation in trials and is working toward providing logistical support to investigators planning and performing large-scale trials.
  • An NINDS workshop (January 2004), "Stroke Risk Assessment and Future Primary Prevention Trials," established the key characteristics of interventions and clinical trial design necessary to proceed with a trial of primary stroke prevention.
  • US Department of Health and Human Services Public Health Action Plan has identified prevention, detection and treatment of risk factors, early detection and treatment of coronary and cerebrovascular disease, and prevention of recurrent cardiovascular events as broad prevention goals for heart disease and stroke prevention under Healthy People 2010.
Unresolved Challenges
  • Very little is known about predictors of long-term patient adherence with stroke prevention therapies. Even less is known about system-wide approaches that are effective at increasing patient adherence to primary and secondary prevention strategies. Furthermore, system-wide and individual interventions to improve physician adherence to recommended therapies have not been systematically evaluated. Thus, major barriers to implementation remain under studied.
  • Prediction of risk of recurrent ischemic events remains poor, both in patients with ischemic stroke and in those with TIA. Blood, genetic, and imaging biomarkers have been understudied and may provide additional prognostic value, as they have in patients with coronary disease.
  • Similarly, prediction of first stroke is very imprecise. Stroke is more heterogeneous than heart disease. Risk stratification tools, such as the Framingham Risk Score and Stroke Prediction Score (from the Cardiovascular Health Study), are limited in their ability to identify those at highest risk for stroke. Clinical factors and biomarkers require further study.
  • Understanding of the acute factors that prompt a stroke or TIA to occur on a given day, at a specific moment, is very limited. Most risk factors are chronic, such as hypertension, hyperlipidemia, and diabetes. Knowledge of the acute predictors of ischemic and hemorrhagic events may help identify new targets for prevention.
  • The appropriate timing of secondary prevention after acute stroke remains uncertain. For example, data on safety of acute lowering of blood pressure in patients with stroke are lacking. Similarly, the safety and benefit of acute antithrombotic therapies with combined antiplatelet agents or in patients with hemorrhagic conversion is unknown. What is the role of stenting in carotid and intracranial atherosclerosis?
  • Prevention of stroke in children remains understudied. The risk of recurrence is very high in this population but even basic questions about antiplatelet therapy have not been addressed.

Priorities and Actions Needed

Priority 1:

Improve implementation of existing, proven stroke prevention guidelines by (1) identifying barriers to such implementation by assessing the individual, healthcare providers, and the healthcare system, (2) studying methods of overcoming these barriers, and (3) supporting the development of research evaluating the effectiveness of innovative initial and recurrent stroke prevention interventions, especially in underserved populations and minority racial/ethnic groups.

Barriers to implementation of proven prevention therapies remain a major public health problem that limits the impact of scientific advancements from clinical trials. For example, anticoagulation for patients with atrial fibrillation, carotid endarterectomy for symptomatic stenosis, and treatment of hypertension are all interventions proven in NIH sponsored randomized trials and implementation of all these therapies has been slow and incomplete. Reasons for this have not been well studied and interventions to improve implementation have rarely been studied systematically.

Embracing research on implementation as part of the mission of the NINDS, inherent in its goal to promote health, is essential to make progress in this area, progress that has been particularly slow in the last 5 years. Other institutions are not well positioned to fund this research because their resources are limited.

Actions Needed

  • Target funding toward implementation research through an RFA/RFP. Although needs in stroke prevention are obvious, such an RFA/RFP could more broadly encompass neurological therapeutics, since implementation in other neurological diseases is also a problem. Research on patient adherence to medications for neurological indications, physician adherence to guideline recommendations, and system-level approaches to improving implementation would all be acceptable areas of research.
  • Convene a summit of potential partners (e.g., CMS, disease-based organizations, health insurers) and academics to identify research synergies in the area of implementation of prevention strategies in adults and children. The goal of the summit would be to produce a plan to accelerate collaborative research in this area, explicitly discussing opportunities for cost sharing and for working together to define appropriate priorities. Support for the development of a national stroke registry should also be considered as part of the agenda; given the expense of this endeavor, rapid evolution in the availability of electronic records, and experience of other institutions in registries (e.g., CDC, CMS, AHA, large healthcare plans), collaboration in promoting a registry is crucial.

Priority 2:

Develop and examine the effectiveness of quantitative risk factor assessment tools that can identify stroke-prone individuals who need aggressive risk factor management and initial and recurrent stroke prevention intervention, with particular emphasis on underserved populations and minority racial/ethnic groups.

There is no standard for stratifying risk of first or recurrent stroke. Furthermore, instruments specific for stroke subtypes (e.g., ischemic stroke or intracerebral hemorrhage) or for use in distinct populations (e.g., children or those with cerebral microbleeds) are lacking. Such instruments, once validated, are likely to be important tools for clinical care, and also for future randomized trials and prospective cohort studies.

Actions Needed

  • Support initiatives that increase access to existing datasets from completed studies through dissemination of de-identified datasets, standardization of variable terms, and through support of retrieval and analysis (as does the NHLBI: see Secondary analyses of existing databases remain an important and efficient source of preliminary evidence that is essential in identifying risk factors for diseases and in deriving risk factor assessment tools. Limiting access to such datasets reduces the potential for new discovery. De-identification is always feasible and is adequate to address concerns about confidentiality, privacy, and HIPAA regulations.
  • Convene a working group to explore development and validation of standardized stroke risk prediction rules (e.g., Framingham risk score) for various settings (e.g., primary prevention, TIA, secondary stroke prevention) and populations (adult and pediatric) to be used in future prevention trials and, possibly for health screening and clinical care. Similar to consensus conferences for developing standard disease definitions or research terms, such a working group may identify preferred tools already developed and may help to set priorities for the research required to establish and validate such tools.
  • Create a repository of biological samples from clinical trials and cohort studies with a standard data format and standardized methods of analysis to facilitate identification of new disease risk factors, for ultimate incorporation in future prediction tools. Such a repository could be virtual (by standardizing and centralizing the data request process but allowing decentralized sample storage by PIs) or real, with a single site used to gather and redistribute samples.

In addition, encourage collection of biological samples as part of clinical trials to allow evaluation of potential stroke risk predictors and modifiers of treatment efficacy. In peer review, ancillary studies in clinical trials are often discouraged. Some process for encouraging submission of such proposals, even from investigators only loosely associated with the principal team, and reviewing them rapidly, before substantial recruitment has occurred, should be developed.

Priority 3:

Support research designed to identify and evaluate innovative stroke prevention treatments and strategies.

Although there has been notable progress in the development of stroke prevention treatments, the overall age-adjusted incidence of stroke has stayed constant during the last 5 years and the burden of stroke has increased as the population continues to age. Progress by industry and the NIH has been at a measured pace and new models for collaboration could accelerate development. Clearly, more highly effective and well tolerated interventions are required.

Actions Needed

  • Pilot test a more rapid mechanism for proposing and reviewing large grants devoted to developing prevention therapies, to increase the focus on promising agents rather than on details of study design and to streamline and accelerate development. Current barriers to writing proposals for large grants, in particular for randomized trials, are high. Proposals are long and detailed, systems for handling all aspects of the study must be developed, and collaborators and potential participating sites must be identified. Few investigators have the necessary expertise to write successful proposals and even fewer can actually get the funded project done based on the plans laid out in the proposal. With a limited number of proposals, funding decisions are not based on the underlying public health import or true scientific and clinical promise of a proposal but rather on the specific interests of the small cadre of qualified investigators. Furthermore, peer review often focuses on details on the proposal, such as the analysis or recruitment plan, rather than on the ultimate impact of the proposed trial. Pre-approval by the NINDS Council is easy to obtain and does not predict well whether a proposal will be ultimately funded.

    A much briefer initial proposal for clinical trials and other large grants should be required. This proposal would detail the background evidence supporting the research question, and the scientific and clinical need and anticipated impact. It would only briefly discuss design issues, such as recruitment and analytical plan. Investigators submitting such proposals might not be clinical trialists and may not have identified a complete team of collaborators. Rather, the proposal would primarily be used to evaluate the potential impact of a trial or other large study. This proposal would be reviewed by peers with input from NINDS staff, including an objective analysis of potential public health impact. With a larger number of such proposals received, it would be expected that most would not be approved for further evaluation. For approved proposals, it would be anticipated that the majority would ultimately be approved for funding. The NINDS staff would work collaboratively with the investigator to identify leadership, statistical support, recruitment plans, and organizational support for the ultimate trial. This new team of investigators-chosen to represent best practices in each area-would then compose a complete grant proposal for the ultimate trial that would be reviewed by peers, now focused on the suitability of the entire study design.

    Such a two-step approval process would allow the NINDS to maintain its tradition of peer review while divorcing questions related to the public health and scientific impact from those related to technical aspects. Furthermore, it would likely increase the total number of proposals received by increasing the pool of people who could propose a trial or other large study, and would increase the quality of ultimate proposals and actual study performance by encouraging collaborations that bring together truly the best representatives of broad areas. Finally, it would create new opportunities to identify potential funding partners for specific proposals, as discussed below.
  • Increase interaction and collaboration between pharmaceutical companies and NINDS to optimize development of new therapeutics and share costs and benefits. This could be accomplished in two ways. First, a conference devoted to development of stroke prevention therapeutics should be convened, bringing together representatives from academia, the NINDS, the FDA, and interested pharmaceutical, biotech, and device companies. Such a conference could work toward establishing standards for pre-clinical testing, set standard definitions and analytical techniques for trials, encourage study and acceptance of appropriately validated surrogate outcomes, identify barriers and potential solutions to common problems such as subject recruitment, and encourage richer collaborations between these disparate entities. The STAIR conference has done this for acute stroke and it has been very influential in the development of new therapies in this area.

    In addition, the NINDS should explore more fluid approaches to negotiating for cost-sharing and access to potential targets with specific companies. Before a decision to fund a trial is made, representatives from the NINDS should meet with any companies that hold patents for studied agents. Ultimate NIH funding of such a trial might be contingent on acceptable negotiation for cost sharing based on an assessment of the potential risks and gains for all parties. If negotiations begin after NIH funding has been assured, motivation to negotiate by the company are severely limited. Such a strategy is only feasible if the decision to fund a trial is based first on a more stringent review of the proposal in light of its clinical and scientific impact in comparison to other potential funding priorities, as proposed above.
  • Develop an institute-initiated research agenda designed to address the most promising interventions in terms of public health benefit, as a supplement to traditional investigator-initiated mechanisms. Based on organized input from medical practitioners and academics with broad representation, the NINDS should identify the most relevant prevention therapies for evaluation in clinical trials. Using the mechanism described above, RFP/RFAs would then be issued to request initial proposals for such studies.
  • Consider a SPOTRIAS-type program focused on primordial, primary, and secondary prevention. Institutions focused on stroke prevention are not necessarily the same ones focused on acute stroke interventions. A network of interventions with a primary interest on prevention could learn from each other and support studies designed at partner institutions. Reducing the barriers to such clinical studies, in particular trials, could accelerate development.
  • Consider supporting research to identify markers of subclinical vascular disease-such as white matter disease, silent infarction, carotid intima-media thickness, and serum biomarkers-and validate them for use as surrogate outcomes for future prevention trials. Surrogate outcomes have greatly accelerated development of therapies in many disease areas, such as for antihypertensives, statins, cancer therapeutics, and even multiple sclerosis. Additional research is required to validate such surrogates for use in clinical trials of stroke prevention therapies. Including these measures in clinical trials could provide the crucial link and validation. Openness to funding such large ancillary studies as part of clinical trials is required.
  • Create an RFP directed at funding prevention research in currently understudied areas, such as pediatric stroke and intracerebral hemorrhage. Industry is not motivated to provide prevention therapies for more rare conditions or for smaller populations. This is clearly part of the mission of NINDS, but such proposals may not fare as well in peer review because of the perceptions that recruitment is a greater problem for more rare diseases and that these questions are not as important as those directed toward more common diseases. An RFP could help provide more balance in the research portfolio.

New Challenges and Opportunities

1. Population-based Primary Stroke Prevention

NINDS has funded several large epidemiological studies geared at promoting an understanding of predisposing risk factors for stroke. The knowledge garnered from these large-scale studies can now be used to identify persons at high risk for first time stroke. Individuals could be selected using multivariable global risk profiles for stroke, such as with the Framingham stroke risk score, and then treated using aggressive evidence-based multimodal therapeutic approaches (pharmacological and non-pharmacological) with a goal to mitigating the incidence of future stroke as well as reducing overall vascular risk.

2. Perinatal Stroke Prevention

Efforts to define the relative contribution of maternal and placental problems as well as fetal and neonatal disorders to the occurrence of perinatal arterial ischemic stroke and sinovenous thrombosis are minimal. Large natural history prospective studies of perinatal stroke risk factors including hematologic and autoimmune disorders, as well as other factors affecting the maternal-fetal environment could be designed with the goal of improving understanding and fostering potential therapeutic avenues (such as the use of antithrombotic agents in prevention) for modifying perinatal stroke risk.

3. Inflammation and Stroke Risk

Several inflammatory biomarkers such as high-sensitivity C-reactive protein (hsCRP), Lipoprotein-associated phospholipase A2 and interleukin-6 have been identified as likely predictors of future stroke risk. However, further studies are needed to properly define the magnitude of these associations beyond other potential confounders (traditional risk factors) that influence vascular risk and to establish causation. Although hydroxymethylglutaryl coenzyme A reductase inhibitors (statins), modulators of the Renin Angiotesnin System, and some antiplatelet agents have been shown to reduce levels of inflammatory markers in addition to their primary mechanisms of action, it is has been difficult to convincingly separate their anti-inflammatory effects from their primary effects with regard to vascular protection. Randomized, controlled trials of agents specifically targeted at lowering only inflammatory markers could provide the strongest evidence of the causal link between inflammation and stroke, as well as provide additional treatment options for those stroke patients at high risk for recurrence who lack conventional stroke risk factors.

4. Personalized Stroke Prevention

Prevailing clinical practice for selecting stroke prevention treatments is generally based on the presence or absence of comorbidities, issues of cost, side effects and other non-genetic characteristics of individual patients. However, recent successes in stroke genomics including the discovery of potential risk factor genes have brought us closer to developing individualized approached to reducing stroke risk. Pharmacogenomics studying the interactions between secondary prevention drugs and gene variants could permit the selection of high-responder versus low-responder populations to various drug regimens in the future. Further studies to determine optimal primary and secondary prevention based on genetic polymorphisms are needed.

5. Obesity, Insulin Resistance and the Metabolic Syndrome

Rapidly rising obesity rates, prospective cohort studies suggesting a link between rising body mass index and stroke, the finding that up to 50% of stroke patients have insulin resistance, and the reported association of the metabolic syndrome with increased risk for cervicocephalic atherosclerosis have made this an important further stroke research. Select CB1 endocannabinoid receptor antagonist and other peroxisome proliferator-activated receptor- (PPAR-) agonists (apart from the glitazones) could be studied with regard to their impact on overall vascular risk reduction in stroke patients.



Co-Chairs: Alexander Dromerick, Elizabeth Noser, Donald Stuss

Members: Bernadette Boden-Albala, John Chae, Leonardo Cohen, Steven Cramer, Nina Dronkers, Pamela Duncan, Susan Fitzpatrick, Tereas Jones, Bryn Kolb, Nicol Korner-Bitensky, Randolph Nudo, Kenneth Ottenbacker, Robert Teasell, Carolee Winstein

NINDS Liaison: Daofen Chen

Statement of the Problem

  • Stroke continues to be the leading cause of long-term disability among adults, and its prevalence will continue to rise as the population ages.
  • Most of the societal impact of stroke has to do with chronic disability, not acute costs. There is a need to focus on chronic and severe impairments and disabilities.
  • Improving function and quality of life are the primary targets of rehabilitation interventions. Depending on the location and extent of brain damage, stroke survivors can have a variety of impairments. Among these are severe motor, sensory, cognitive, communicative, executive, emotional, mood, and social problems. It is well established that any one of these can substantially impact quality of life, not only in patients, but also in caregivers.
  • Basic and translational science of recovery and rehabilitation remains largely unexplored and ignored by most biotechnology and pharmaceutical companies.

Since 2001, there have been three levels of success: seminal advances, significant scientific findings, and important advances in scientific methods and theories necessary to advance the science of stroke rehabilitation. These scientific advances are all linked to SPRG recommendations. Each has raised unresolved key scientific questions and provides the impetus for future priorities that will impact stroke recovery.

Seminal Advances

A. Funding and execution of the first large scale Phase III Randomized Controlled Trials of Rehabilitation interventions: Two major NIH-funded Phase III trials of motor rehabilitation interventions have been completed, CIM Therapy in stroke (Wolf et al) and Treadmill training in SCI (Dobkin et al). Two other large scale multicenter trials funded by the NIH are underway, including treadmill training in stroke (Duncan et al) and amphetamine in stroke (Goldstein et al). Similar efforts are underway in Canada and around the world.

Key unresolved scientific questions:

1. Establishing "pharmacokinetics" of training: Dosing, timing of administration, half-life, does training paradigm matter?

2. Effect of clinical parameters: Lesion location, prior lesions, baseline severity, coexisting cognitive impairments

3. Rehab/Recovery trial methodology: What is a clinically significant effect? How do measures of deficits relate to real world function?

4. What is most effective way to do Phase II development of behavioral interventions?

Newly emerged key research areas:

1. Differential impact of varying lesion locations on responsiveness to restorative interventions.

2. How pretreatment brain physiology (measured using fMRI, TMS or other physiological methods) relate to treatment responsiveness.

3. Clinical trials of emerging potential treatments for language and cognitive impairment are necessary.

B. Improved animal models of stroke and recovery: Several studies in rodent and non-human primate models have now documented improved motor recovery and increased neuroplasticity by introducing rehabilitative training protocols early rather than late after experimental ischemic stroke. These new findings suggest that moderate behavioral training results in improved recovery when introduced early (by one week) as compared with late (one month) after infarct. These findings have important implications for the timing of rehabilitative protocols in human stroke.

While rodent models of stroke recovery have been utilized for decades, one of the drawbacks has been that recovery was typically rapid, and chronic deficits were rarely documented. However, the recent application of behavioral methodology to examine skilled forelimb reaching and grasping in rodent models has allowed the development of outcome measures with much greater sensitivity. This technological advance has allowed the demonstration of chronic deficits in rodent models after stroke, and thus the testing of rehabilitative behavioral protocols that require several weeks to evaluate. This advance will greatly enhance our ability to test the efficacy of new interventions for recovery in chronic stroke.

Key unresolved scientific questions:

1. What is the ecological validity of animal models of motor recovery? Issues similar to acute stroke models: age, comorbidities, lesions, etc. There is a need for better refinement and validation of aged animal models of stroke that are suitable for investigations of behavioral and brain changes in the chronic period.

2. Subcortical stroke models

3. Modeling cognitive rehabilitation

Newly emerged key research areas:

1. What is different about the brain at key time points of responsiveness?

2. Chronic period: What is the progression of brain changes even in the absence of overt behavioral training?

3. Can early plasticity state be reconstituted later, or prolonged?

Significant Scientific Accomplishments

A. Substantial increase in the understanding of how training parameters and subject characteristics change response to motor rehabilitation: In particular, task specific training has emerged as the dominant approach to motor restoration. Other findings from the motor learning field including the delivery schedule of training, motivation and contextual factors are now under active study. Characteristics of treatment participants, including severity of motor deficits and co-existing sensory or cognitive impairments are also under investigation.

Key unresolved scientific questions:

1. Factors that affect responsiveness to rehabilitation therapies - age, co-morbidities, gender, race/ethnicity, socioeconomic status, and the physiological and clinical features of stroke -- are not well understood.

Newly emerged key research areas:

1. Define attributes of physical, cognitive, and communicative rehabilitation interventions that promote recovery/compensation.

B. Better understanding of the anatomic and functional relationships among brain structures that support cognition has lead to more rational approaches to cognitive rehabilitation: Functional imaging and state of the art structure -function studies are yielding new insights into the fundamental processes that underlie cognition and behavior in normals and in persons with stroke.

This advance has led to the development of theory-driven interventions; well-designed small scale trials have been completed or are now underway.

Key unresolved scientific questions:

1. Does cognitive intervention targeted to one specific cognitive deficit actually have long lasting (and real-life) benefits, or must such interventions be completed within a more comprehensive approach?

2. To what extent do patients exhibit multiple deficits in cognition and/or language and to what extent do these interact?

3. Are there aspects of cognition and/or language that seem to rely on the same brain areas, thus implicating these regions for interactive treatment approaches?

4. Are there lesioned areas that uniquely affect different aspects of language and/or cognition such that spared areas and their functions could be "tapped" during rehabilitation and recovery?

Newly emerged key research areas:

1. What is the interaction of cognitive deficits with language disorders?

2. Behavioral and psychosocial studies are needed to promote cognitive rehabilitation and psychosocial support, understand the role of environmental factors, increase participation, and improve quality of life.

Important Methodological and Theoretical Advances

A. Improved experimental techniques have been applied to rehabilitation research. Advances in several methods are pivotal to future development in stroke treatment research. For example, a) fMRI has become a widely used tool to better understand the physiology of recovery, specifically how brain activation distant from the lesion relates to recovery. b) New imaging analysis techniques provide the unique opportunity to evaluate whole brain system re-organization. c) Motor learning studies are now being incorporated into rehabilitation treatments undergoing testing.

Key unresolved scientific questions:

1. What is the interaction between lesion location and extent, and how do these map onto functional re-organization? How do functional imaging changes interact with the behavioral changes after focal stroke?

Newly emerged key research areas:

1. Utilizing biomarkers (fMRI, ERPs, DTI, etc.) to mark the neurobiological changes that occur during recovery and rehabilitation and verify that changes have occurred.

B. The disability associated with vascular cognitive impairment and other stroke syndromes with subtle cognitive impairment is now an area of active research. VCI is an extremely common condition, and may yield important insights into other clinically important impairments such as MCI and mild traumatic brain injury.

Key unresolved scientific questions:

1. What are the effects of milder cognitive impairments on patients' mood, as well as their ability to cope, return to work, and function in social situations. (Clinically, we see that patients with milder impairments often present as more depressed and frustrated because caregivers and co-workers do not appreciate the effects of these mild impairments. They may be walking and talking, but they cannot function normally, as they once did.)

Newly emerged key research areas: We need to understand interactions among various cognitive impairments.

1. How do cognitive impairments affect each other.

2. How does cognitive impairment affect response to treatment-motor, language, ADL, etc?

3. Understand the interaction of cognitive deficits with language disorders.

C. New scientific knowledge from other areas provides the basis to develop directed cognitive, emotional and psychosocial interventions. For example, different "executive" functions have now been demonstrated to be related to specific frontal regions, suggesting targeted interventions for specific disorders.

Key unresolved scientific questions:

1. What are the interactions among different cognitive systems (e.g., "executive" to language) and between cognitive and motor systems (can attentional capabilities be used in motor recovery?)

Newly emerged key research areas: We have seen developments in the area of lesion-symptom mapping of mathematical functions, problem-solving, aspects of speech production and language comprehension, among other areas of cognition and language. Recent work with grid electrode placement in neurosurgical cases can now look at the time course during which these areas are recruited as well as how they interact to form networks that underlie complex cognitive processes in humans. Understanding these brain-behavior relationships has been critical in advancing neurocognitive rehabilitation techniques that can target the specific cognitive functions affected by region-specific lesions. Similarly, spared regions and their related cognitive functions can be accessed during rehabilitation to facilitate the reorganization of function and the development of compensatory strategies.

Future Proposed Priorities

Priority 1: Neurobiology of recovery

  • Cellular and molecular mechanisms of plasticity
  • Prolonging or reconstituting plastic period
  • How does pre-treatment brain physiology relate to treatment responsiveness?
    • fMRI, TMS, MEG, etc
  • Deeper understanding of anatomic and functional networks that support motor, sensory & cognitive systems
    • Subcortical/brainstem/spinal cord motor centers
    • Networks for language/attention/exec function
    • Sensory systems
  • Genetic modifiers of recovery

Priority 2: Clinical Trials

  • Subject key aspects of currently used rehabilitation strategies to pivotal RCT's
    • Timing, dosage, half life of intervention
    • Training paradigm effects
  • Link neurologic deficits with disability and social participation
    • What deficits disable, and at what severity?
    • What are clinically important improvements in deficits?
    • How do different deficits interact to worsen disability?
    • Can targeting a single deficit lead to significant improvement?
  • Refine and accelerate Phase II development path for emerging behavioral therapies
    • Testing behavioral interventions is different in key respects from drug trials
  • Effects of modifying variables
    • Lesion location, prior lesions, coexisting cognitive impairments, medical comorbidities
  • Development of standardized treatment and measures for recovery agent trials

Proposed Future Priority 3

  • Treatments for persons with chronic or severe impairments and disabilities*
    • Compensation strategies:
      • Targeting cardiovascular, muscle systems
      • Assistive technology
      • Environment
    • Restorative strategies:
      • Reconstituting plastic state
      • RCT's of training paradigms in chronic stroke

Proposed Future Priority 4

  • Cognitive Rehabilitation
    • Aphasia
      • Phase II trials
      • Interaction with other cognitive deficits
      • Relationship of aphasia characteristics with community reintegration
    • Subtle cognitive impairment
      • Executive function and reintegration
    • Post-stroke depression
      • Pathophysiology, treatment

Other proposed future priorities

  • Stroke sequelae in working age
  • Engineering-based restorative approaches:
    • Cortical prosthetics
    • Cognitive prosthetics
    • In-home monitoring for cognitively disabled
    • Exoskeletal robotic orthotics, bions, etc
  • Define clinical indications for and outcomes from different levels of rehab care.



Co-chairs: Sandra Black, Philip Gorelick, Steven Greenberg

Members: Charles DeCarli, Martin Dichgans, Frank Faraci, Antoine Hakim, David Nyenhuis, Reinhold Schmidt, William Van Nostrand

NINDS Liaison: Gabrielle Lablanc

Scientific Advances in Vascular Cognitive Impairment (VCI) since 2001

  • Subcortical Small Vessel Disease Consensus has emerged around subcortical small vessel disease-much of it clinically asymptomatic-as the major cerebrovascular contributor to cognitive impairment in the elderly. Silent small vessel brain disease is a silent epidemic.

Population-based clinical-radiographic studies (Rotterdam Scan, Cardiovascular Health, Framingham, Austrian Stroke Prevention) have demonstrated small vessel disease, silent brain infarction, and white matter signal abnormalities to be highly prevalent and potent risks for cognitive impairment in the elderly. The prevalence of silent brain infarction in unselected population-based subjects has been variously estimated at 11% for subjects age 55 to 70, 21% for subjects age 60 to 90, and 23% for subjects age ?65. The incidence of new silent infarction is equally striking, estimated at 13% over 3.4 years and 18% over 5 years in longitudinal population-based studies of the elderly. Though termed "silent," these subclinical infarctions are clearly associated with substantial risk for cognitive decline and symptomatic stroke. Subjects with silent brain infarction at baseline had more than two-fold increased risk of incident dementia and significantly greater overall cognitive decline in the Rotterdam Scan study.

The close association between small infarctions and cognitive impairment is consistent with a large body of data demonstrating a synergistic relationship between Alzheimer's disease (AD) pathology and vascular risk factors/disease in producing dementia. Coexistent Alzheimer's and cerebrovascular disease is emerging as the most common pathological substrate of dementia based on multiple population-based clinical-pathological studies (Nun, VA Puget Sound, Medical Research Council). It remains unclear whether cerebrovascular disease directly potentiates AD pathology or if the two processes affect cognition independently and synergistically.

  • Imaging of Small Vessel Disease There have been substantial technical advances in detecting the presence and progression of small vessel brain disease and its interaction with brain atrophy and cognitive decline.

The white matter T2-hyperintensities (WMH) sensitively detected by MRI scan have emerged as an important marker of the presence and progression of small vessel brain disease. Recent data indicate that progression of WMH parallels cognitive decline, supporting the clinical importance of this radiographic finding. In the Cardiovascular Health Study, for example, lesion progression was associated with decline in MMSE and Digit Symbol Substitution Test. Similarly, the PROSPER study found that increases in periventricular WMH associated with declines in mental processing speed, while the Austrian Stroke Prevention Study found that cognitive decline related directly to loss of brain substance occurring together with progression of WMH burden. The Austrian data emphasize important interactions between progression of subcortical ischemic brain damage and concomitant brain atrophy when evaluating the morphological substrates of cognitive decline in elderly subjects. Notably, the interaction between white matter lesions and brain atrophy in their effects on cognition also exists in patients with cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL), a relatively pure form of subcortical VCI without associated AD pathology.

Progression of WMH is common, even when detected by relatively insensitive visual methods. The Austrian Stroke Prevention Study, for example, reported WMH progression over 3 years in 17.9% of a community-dwelling cohort of middle-aged and elderly subjects free of signs or symptoms of neuropsychiatric disease, with 8.1% of participants having marked lesion progression. In the Cardiovascular Health Study, 28% of individuals had WMH increases, mostly by 1 grade, after a five year follow-up period, while the Rotterdam Scan Study reported a 27% frequency of progression (9% with marked progression) after a mean follow-up of 3.3 years.

Volumetric measurement has more recently emerged as the method of choice to evaluate longitudinal changes in WMH. Volumetric methods vary from fully manual outline techniques that are labor-intensive and modestly accurate, to fully automated methods not requiring user interaction and not yet fully validated. Semi-automated algorithms requiring some corrective user interaction are currently thought to be the most practical methods to measure change in white matter lesion load, as processing time is shorter and intra-and inter-rater reliability is higher than in manual contouring. These methods may be used in medium-sized trials, but remain labor-intensive for large trials. Automated image post-processing techniques have also been developed to study global brain parenchymal loss over time. One method that has been proposed as a potential surrogate marker is the boundary shift integral approach, a semi-automated method involving subtraction of follow-up images from co-registered baseline images to calculate the positional shift in tissue boundaries and thus the amount of tissue loss. Another widely used method is SIENA (Structural Image Evaluation, using Normalization of Atrophy), which has the advantage of being relatively insensitive to differences in scanning parameters.

Newer MRI techniques have the potential to characterize and quantify the amount of tissue damage within and outside of conventionally imaged lesions. These techniques include diffusion-tensor imaging, magnetization-transfer imaging, and magnetic resonance spectroscopy. Each has shown promise as a surrogate marker of disease progression in treatment trials, with diffusion-tensor imaging in particular demonstrating progression over time and correlation with clinical findings in a study of CADASIL patients. One challenge in applying these imaging techniques to multi-center studies is that the results are highly dependent on pulse sequence parameters.

  • Neuropsychological Test Batteries Consensus has emerged around the NINDS-CNS Harmonization Working Group cognitive testing protocols as standardized batteries for measuring impairment in domains vulnerable to large and small vessel brain injury.

Defining a battery of standardized tests for identifying and quantifying the cognitive impairments associated with small vessel brain disease has been a major goal in the VCI field, particularly with regard to the design of clinical trials. Testing instruments in current use have been largely designed for AD and are relatively poor in sampling executive dysfunction. Based on these considerations, the NINDS and the Canadian Stroke Network convened a Harmonization Workshop in April 2005. The goal of the workshop was to define a set of data elements to be collected in future studies to more fully define VCI as a disease entity, understand its etiology and identify targets for treatment. Working groups were formed in neuropsychology, neuroimaging, epidemiology, pathology, and experimental research. The Neuropsychology Working Group was charged with recommending test protocols that could characterize VCI in multi-center investigations. Because different protocols serve different purposes, the group produced three protocols, one that required approximately 60 minutes, a second that required 30 minutes and a third that required five minutes. It was envisioned that the 60 Minute Protocol be used in studies that require a breakdown of cognitive abilities by domain, so the protocol contains recommended tests in four domains: executive/activation, language, visuospatial, and memory. In addition, tests were selected to examine neurobehavioral change and mood. Tests for the 30 Minute Protocol were selected from the 60 Minute Protocol to be used as a clinical screening instrument for patients with suspected VCI. Finally, a Five Minute Protocol, based largely on the Montreal Cognitive Assessment, was devised for possible use by primary care physicians, nurses and other allied health professionals who need a quick screening in their office or at the bedside. The Five Minute Protocol was also designed for very large epidemiologic studies or clinical trials in which sensitivity and ease of administration are especially important. In addition, the Five Minute Protocol was designed so that, once validated, it could be administered by telephone.

Further work in this area is still needed to examine the reliability, validity, and usefulness of these protocols in multi-center clinical trial settings.

  • Small Vessel Physiology and Biomarkers Studies of small vessel physiology in animals and biomarkers in humans have yielded substantial advances in identifying pathogenic pathways involved in small vessel injury.

Studies of cerebral small vessel physiology in anesthetized animals have identified common mechanisms that may underlie cerebral vascular abnormalities in multiple diseases. Oxidative stress and inflammation, for example, appear to play major roles in various vascular disease states. The renin-angiotensin system is another candidate pathway; in addition to its well-recognized role in hypertension, this system may contribute to vascular abnormalities in atherosclerosis, diabetes/insulin resistance, hyperhomocysteinemia, and aging. The pleiotropic role of these pathways may help explain the observation from clinical trials that agents such as angiotensin converting enzyme-inhibitors and statins appear to reduce risk of stroke beyond their effects on blood pressure or serum lipids alone.

Complementing these studies of extracranial and leptomeningeal vessel physiology is progress in developing approaches to studying the small intraparenchymal vessels and their role in the neurovascular unit (NVU). Experiments using in vivo imaging of brain slices have begun to explore the impact of ischemia on the NVU as well as the role of the NVU in the coupling between neuronal activity and blood flow. Another emerging area with therapeutic promise is the study of molecular mechanisms that protect against vascular oxidative stress. Recent data have implicated endogenous transcription factors and pathways regulating entire clusters of genes or proteins in mediating vascular protection at multiple levels.

Among candidate biomarkers, C-reactive protein and homocysteine have already been implicated in VCI by population-based longitudinal studies. With accelerated progress in proteomics and large-scale biochemical screening methods, others are rapidly emerging. For example, correlations between CSF concentrations of TNF-?, a proinflammatory cytokine mediating myelin damage, and sulfatide, a marker of white matter degradation, suggest that this apoptosis-inducing cytokine may lead to the death of oligodendrocytes in contributing to white matter degeneration. Similarly, CSF IL-6 levels are increased in patients with vascular dementia, suggesting that inflammatory mechanisms may be involved in the development of cognitive decline in some patients with cerebrovascular disease.

Biomarkers for the ß-amyloid peptide (Aß) are a growing focus of VCI studies because of this peptide's role as primary constituent of both the senile plaques of AD and the small vessel deposits of cerebral amyloid angiopathy (CAA). CSF measures of Aß and the microtubule-associated protein tau are well established as AD biomarkers and may have utility in CAA as well. An intriguing correlation has more recently emerged between plasma Aß and the presence on MRI of lacunar infarction and WMH. The mechanism for this association, noted initially in the Rotterdam Scan study and largely confirmed in a population of AD and CAA subjects, remains to be determined. The development of Pittsburgh compound B (PiB), a lipophilic Aß-binding compound, offers potentially the most powerful biomarker for Aß. PiB-PET imaging has been shown to detect AD pathology and possibly CAA pathology as well. Addition of PiB-PET imaging to longitudinal studies of cognition offers the exciting possibility of teasing apart the contributions of AD and small vessel disease in generating cognitive impairment.

  • Transgenic Mouse Models of Human Small Vessel Disease The development of transgenic mouse models for specific human small vessel pathologies offer potential vehicles both for studying cerebrovascular dysfunction and for testing candidate approaches to treatment.

Investigators have generated transgenic mice that demonstrate many of the hallmark features of CADASIL and CAA. CADASIL transgenic mice carrying a typical Notch3 mutation (TGNotch3R90C) recapitulate the vascular pathology of CADASIL, including the deposition of Notch3-immunoreactive material and ultrastructural deposits in small arteries and precapillaries. These mice also develop characteristic degenerative vascular changes such as smooth muscle cells loss. Hemodynamic studies in R90C transgenic mice have revealed disturbed autoregulation and impaired vasomotor response to hypercapnia and acetazolamide, which may contribute to ischemia. It is notable, however, that there are no parenchymal lesions even in very old TGR90C mice. Additional models with stronger or inducible promoters, other mutations, and other genetic backgrounds may be needed to further explore the mechanisms of ischemia in this condition and in small vessel disease in general.

Several human transgenic mouse models with some degree of CAA have been generated by expression of mutant forms of the ß-amyloid precursor protein (ßAPP). The first mouse model reported to develop appreciable Aß-type CAA was the APP23 transgenic mouse expressing ßAPP with the Swedish double familial AD mutation under the control of the neuron-specific murine Thy-1 promoter. Aged APP23 mice develop parenchymal amyloid plaques and cerebrovascular amyloid as well as some of the key pathological features of human CAA such as vascular-associated inflammation, vascular cell death, and microhemorrhage. Advanced CAA with associated inflammation, vascular cell death and microhemorrhage has also been noted in aged Tg2576 mice. Both of these mouse lines rely on overexpression (approximately 6 to 7-fold) of ßAPP, generating high amounts of wild-type Aß. A more recent alternative strategy has been to utilize forms of ßAPP that yield mutant Aß peptides associated with familial CAA. APPDutch mice express human ßAPP harboring the Dutch familial CAA mutation under the control of the Thy-1 promoter. These mice are unique in that they develop CAA in the absence of parenchymal amyloid plaques, demonstrating extensive Aß deposits in leptomeningeal vessels and cortical vessels, neuroinflammation, and microhemorrhages at advanced ages (?24 months). The Tg-SwDI mouse represents a new model that expresses human ßAPP harboring the Swedish familial AD mutation in addition to both the Dutch and Iowa familial CAA mutations under control of Thy1. Unique to this model, the transgene encoded human ßAPP is expressed at levels below that of murine ßAPP. Tg-SwDI mice nevertheless show early-onset cerebral small vessel amyloidosis in the form of fibrillar perivascular/vascular Aß deposits starting at the age of 4 to 5 months. A strong localized neuroinflammatory response is observed restricted to the affected microvessels, while hemorrhages are scarce.

A recent series of studies identified a mutation in the procollagen type IV alpha 1 (COL4A1) gene as a novel cause of small vessel brain disease. Though initially identified in a transgenic mouse line by random mutagenesis, mutations in COL4A1 were subsequently shown to cause hemorrhagic strokes and white matter lesions in several unrelated human families. This observation highlights the potential of transgenic mouse studies to identify highly unexpected molecular pathways associated with cerebrovascular disease.

Unresolved Challenges

  • Lack of Definitive Animal Model for Human VCI There is still no ideal model system for screening and identifying candidate pharmacologic approaches to prevention of VCI.

Despite the exciting advances in physiologic and transgenic animal models outlined above, none has been clearly validated as a faithful model of all aspects of human VCI. Hypertensive microvasculopathy in particular, presumably the most common form of cerebral small vessel disease, has no definitive model system. Two recent mouse models of cerebrovascular hypertension-a line of transgenic mice that overexpress human renin and angiotensin and wild-type mice with surgical coarctation of the transverse aorta-have been shown to develop cerebrovascular pathology, although neither has yet been found to mimic human small vessel disease. One barrier to mouse modeling of VCI is the animal's low volume of white matter; the alternative approach of using non-human primates is feasible, but technically difficult because of expense, the high level of required experimental expertise, and the lesser flexibility for genetic manipulation. Another barrier to animal modeling of VCI is that most studies are performed in younger animals, whereas VCI is primarily a disease of the aging brain. Despite the fact that aging has an enormous negative impact on the cerebral circulation, we know very little about the underlying changes in biology associated with aging per se.

  • Need for Randomized Intervention Trials Non-randomized observational trials of candidate preventive treatments for VCI are difficult to interpret because of the likelihood of confounding.

Observational studies have been the source for much of what we have learned about VCI and are an attractive method for identifying behaviors, diets, medications, or other interventions associated with the presence or absence of small vessel disease. Recent experience with observational findings that could not be replicated in randomized trials, however, highlights the potential for confounding even in large, well-designed observational cohorts. We therefore emphasize the importance of randomized trials as a step towards identifying candidate interventions.

  • Challenges of Large-Scale Randomized Intervention Trials Interventional trials aimed at primary prevention of VCI are challenging to design and execute.

One complicating factor in designing randomized trials for prevention of VCI is the large sample size and long duration of treatment and follow-up that may be required to detect effective prevention. Sample size and follow-up can likely be reduced by choosing high-risk populations, although halting already-established small vessel disease may prove difficult. Focusing on high-risk populations also highlights several other challenges in study design, including the fact that even those high-risk subjects randomized to "placebo" still need to receive treatment for their vascular risk factors, as well as concerns that the screening tests themselves (e.g. screening MRI scans) may alter patient or physician behavior.

Priorities and Actions Needed

  • Experimental Systems Modeling Human Small Vessel Disease and VCI We recommend continued aggressive development and validation of experimental systems modeling human small vessel disease and VCI.

A major goal remains developing systems that allow identification of key molecular pathways in small vessel brain injury and screening of candidate therapies. Development of a faithful model of hypertensive microvasculopathy would be an extremely useful addition to the spectrum of available experimental systems. Another promising direction for experimental research will be use of existing models systems to explore the interaction between various risk factors, e.g. the relationship between blood pressure and CAA, CADASIL, and normal aging in promoting small vessel injury.

  • Human Studies to Define High-Risk Populations, Validate Imaging Endpoints, Define Biomarkers, and Identify Candidate Treatments for VCI We recommend expanded use of small- to moderate-scale human studies to 1) define populations at high risk for VCI progression, 2) validate clinically meaningful radiographic endpoints, 3) explore potential biomarkers, and 4) identify candidate treatments. The overall goal of these studies is to lay the groundwork for definitive trials of candidate prevention strategies.

Several existing trials highlight the potential utility of these types of studies. The Austrian Stroke Prevention Study, for example, explored risk factors for progression of WMH volume. Change in WMH volume for the whole group was small (median 0 and 0.1 cm3 increase at 3- and 6-year follow-up respectively). Stratification by baseline grade of WMH, however, demonstrated that the small overall increase in lesion volume was due to the negligible rate of lesion progression in the large subset of participants with no or only punctate abnormalities. In contrast, study participants with a baseline finding of early confluent or confluent changes showed a remarkably rapid increase in lesion volume (median volume increase at 6 years of 2.7 cm3 for early confluent and 9.3 cm3 for confluent).

The potential of small- to moderate-scale studies using surrogate markers of outcome to identify promising candidate treatments was recently demonstrated by a substudy of the PROGRESS trial. In this substudy of 192 PROGRESS participants, subjects randomized to perindopril plus indapamide demonstrated relatively modest reductions in blood pressure (11.2/4.3 mm Hg in systolic/diastolic blood pressure). WMH progression was nevertheless substantially reduced (mean volume of incident WMH 0.4 mm3 for treated versus 2.0 mm3 for placebo) over a mean 36-month follow-up. Notably, the protective effect of treatment was greatest in subjects with the most advanced baseline WMH burden, offering evidence that interventions may indeed succeed even in fully established small vessel disease.

As support builds for the validity and efficiency of WMH as a surrogate marker of VCI, we recommend that measurement of WMH volume at baseline and follow-up be performed routinely as part of small- to moderate-scale studies. Use of standardized MRI acquisition protocols such as those recommended by the NINDS-CNS Harmonization Working Group would facilitate multi-center studies and comparisons.

  • Long-Term Prevention Trials We recommend long-term prevention trials for VCI and other forms of cognitive impairment associated with atherosclerotic risk factors.

These trials may be established as "add-on" or "piggy-back" studies to existent, large-scale trials or as primary studies. In either case, a crucial consideration will be use of validated VCI test batteries rather than insensitive measures such as the MMSE. Because patients with atherosclerotic risk factors are considered at high risk for stroke and cardiovascular disease, treatment of these factors will be indicated according to appropriate clinical guidelines. VCI and other atherosclerotic trial treatment paradigms will therefore likely require randomization to 'usual care' versus 'intensive- or stepped-level care'. Large, simple trial designs are to be encouraged to conserve financial resources, enhance recruitment of large, representative study populations, and allow more worldwide study participation.

New Challenges and Opportunities

  • VCI, unlike AD, may be preventable with currently available therapies. Our knowledge of vascular risk factors and pathophysiology should make it possible to identify effective vasculoprotective approaches for prevention of VCI-sooner than we will be able to identify anti-Aß agents for prevention of AD.
  • The basic pathophysiological mechanisms underlying various cerebrovascular diseases may be shared (e.g., inflammation, oxidative stress, apoptosis). Consideration should therefore be given to shifting the prevention paradigm upstream to identify and intervene on common pathophysiological pathways leading to downstream disease and cognitive impairment in the at-risk brain.
  • AD may be potentiated or triggered by atherosclerotic risk factors or disease. The borderline between VCI and AD needs to be better defined and forms of AD driven by atherosclerotic risk factors or disease need to be identified and considered for appropriate interventional study.
  • Carefully designed and performed clinical studies in VCI will have the collateral benefit of providing a wealth of data on the molecular biology and genetics of this disorder.
  • The high population-wide attributable risk associated with "silent" small vessel disease makes it an extremely attractive target for primary prevention.

Last Modified October 18, 2015