This Article Abstract Full Text (PDF) Correction (v34,p2774) All Versions of this Article: 34/8/e109    most recent 01.STR.0000082721.62796.09v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Higashida, R. T. Articles by Furlan, A. J. Search for Related Content PubMed PubMed Citation Articles by Higashida, R. T. Articles by Furlan, A. J. Related Collections Anticoagulants Thrombolysis Arterial thrombosis CT and MRI Acute Cerebral Infarction Acute Stroke Syndromes Embolic stroke Emergency treatment of Stroke Angiography Doppler ultrasound, Transcranial Doppler etc. Pathology of Stroke Fibrinolysis Risk Factors for Stroke Related Article

(Stroke. 2003;34:e109.)
© 2003 American Heart Association, Inc.

Original Contributions

Trial Design and Reporting Standards for Intra-Arterial Cerebral Thrombolysis for Acute Ischemic Stroke

Randall T. Higashida, MD Anthony J. Furlan, MD for the Technology Assessment Committees of the American Society of Interventional and Therapeutic Neuroradiology and the Society of Interventional Radiology

For the Technology Assessment Committees (Committee members Heidi Roberts, Thomas Tomsick, Buddy Connors, John Barr, William Dillon, Steven Warach, Joseph Broderick, Barbara Tilley, and David Sacks) of the American Society of Interventional and Therapeutic Neuroradiology and the Society of Interventional Radiology.

Correspondence to Randall T. Higashida, MD, University of California, San Francisco Medical Center, Departments of Radiology, Neurological Surgery, Neurology, and Anesthesiology, Division of Interventional Neurovascular Radiology, 505 Parnassus Ave, Rm L-352, San Francisco, CA 94143-0628. E-mail Randall.Higashida{at}radiology.ucsf.edu

	Abstract

Top Abstract Introduction IV Trials in Acute... IA Trials in Acute... Limitations of IA Thrombolysis Predictors for Treatment Success Pretreatment Evaluation Defining Patient Selection Defining Outcome of Therapy Treatment Description Posttreatment Evaluation:... Complications and Adverse Events Comparison Between Treatment... Costs Conclusions References

 
Background and Purpose— The National Institutes of Health (NIH) estimates that stroke costs now exceed $45 billion per year. Stroke is the third leading cause of death and one of the leading causes of adult disability in North America, Europe, and Asia. A number of well-designed randomized stroke trials and case series have now been reported in the literature to evaluate the safety and efficacy of thrombolytic therapy for the treatment of acute ischemic stroke. These stroke trials have included intravenous studies, intra-arterial studies, and combinations of both, as well as use of mechanical devices for removal of thromboemboli and of neuroprotectant drugs, alone or in combination with thrombolytic therapy. At this time, the only therapy demonstrated to improve outcomes from an acute stroke is thrombolysis of the clot responsible for the ischemic event.

There is room for improvement in stroke lysis studies. Divergent criteria, with disparate reporting standards and definitions, have made direct comparisons between stroke trials difficult to compare and contrast in terms of overall patient outcomes and efficacy of treatment. There is a need for more uniform definitions of multiple variables such as collateral flow, degree of recanalization, assessment of perfusion, and infarct size.

In addition, there are multiple unanswered questions that require further investigation, in particular, questions as to which patients are best treated with thrombolysis. One of the most important predictors of clinical success is time to treatment, with early treatment of <3 hours for intravenous tissue plasminogen activator and <6 hours for intra-arterial thrombolysis demonstrating significant improvement in terms of 90-day clinical outcome and reduced cerebral hemorrhage. It is possible that improved imaging that identifies the ischemic penumbra and distinguishes it from irreversibly infarcted tissue will more accurately select patients for therapy than duration of symptoms. There are additional problems in the assessment of patients eligible for thrombolysis. These include being able to predict whether a particular site of occlusion can be successfully revascularized, predict an individual patient’s prognosis and outcome after revascularization, and in particular, to predict the development of intracerebral hemorrhage, with and without clinical deterioration. It is not clear to assume that achieving immediate flow restoration due to thrombolytic therapy implies clinical success and improved outcome. There is no simple correlation between recanalization and observed clinical benefit in all ischemic stroke patients, because other interactive variables, such as collateral circulation, the ischemic penumbra, lesion location and extent, time to treatment, and hemorrhagic conversion, are all interrelated to outcome.

Methods— This article was written under the auspices of the Technology Assessment Committees for both the American Society of Interventional and Therapeutic Neuroradiology and the Society of Interventional Radiology. The purpose of this document is to provide guidance for the ongoing study design of trials of intra-arterial cerebral thrombolysis in acute ischemic stroke. It serves as a background for the intra-arterial thrombolytic trials in North America and Europe, discusses limitations of thrombolytic therapy, defines predictors for success, and offers the rationale for the different considerations that might be important during the design of a clinical trial for intra-arterial thrombolysis in acute stroke. Included in this guidance document are suggestions for uniform reporting standards for such trials. These definitions and standards are mainly intended for research trials; however, they should also be helpful in clinical practice and applicable to all publications.

This article serves to standardize reporting terminology and includes pretreatment assessment, neurologic evaluation with the NIH Stroke Scale score, imaging evaluation, occlusion sites, perfusion grades, follow-up imaging studies, and neurologic assessments. Moreover, previously used and established definitions for patient selection, outcome assessment, and data analysis are provided, with some possible variations on specific end points. This document is therefore targeted to help an investigator to critically review the scales and scores used previously in stroke trials.

This article also seeks to standardize patient selection for treatment based on neurologic condition at presentation, baseline imaging studies, and utilization of standardized inclusion/exclusion criteria. It defines outcomes from therapy in phase I, II, and III studies. Statistical approaches are presented for analyzing outcomes from prospective, randomized trials with both primary and secondary variable analysis. A discussion on techniques for angiography, intra-arterial thrombolysis, anticoagulation, adjuvant therapy, and patient management after therapy is given, as well as recommendations for posttreatment evaluation, duration of follow-up, and reporting of disability outcomes.

Imaging assessment before and after treatment is given. In the past, noncontrast CT brain scans were used as the initial screening examination of choice to exclude cerebral hemorrhage. However, it is now possible to quantify the volume of early infarct by using contiguous, discrete (nonhelical) images of 5 mm. In addition, CT angiography by helical scanning and 100 mL of intravenous contrast agent can be used expeditiously to obtain excellent vascular anatomy, define the occlusion site, obtain 2D and 3D reformatted vascular images, grade collateral blood flow, and perform tissue-perfusion studies to define transit times of a contrast bolus through specific tissue beds and regions of interest in the brain. Dynamic CT perfusion scans to assess the whole dynamics of a contrast agent transit curve can now be routinely obtained at many hospitals involved in these studies. The rationale, current status of this technology, and potential use in future clinical trials are given.

Many hospitals are also performing MR brain studies at baseline in addition to, or instead of, CT scans. MRI has a high sensitivity and specificity for the diagnosis of ischemic stroke in the first several hours from symptom onset, identifies arterial occlusions, and characterizes ischemic pathology noninvasively. Case series have demonstrated and characterized the early detection of intraparenchymal hemorrhage and subarachnoid hemorrhage by MRI. Echo planar images, used for diffusion MRI and, in particular, perfusion MRI are inherently sensitive for the susceptibility changes caused by intraparenchymal blood products. Consequently, MRI has replaced CT to rule out acute hemorrhage in some centers. The rationale and the potential uses of MR scanning are provided.

In addition to established criteria, technology is continuously evolving, and imaging techniques have been introduced that offer new insights into the pathophysiology of acute ischemic stroke. For example, a better patient stratification might be possible if CT and/or MRI brain scans are used not only as exclusion criteria but also to provide individual inclusion and exclusion criteria based on tissue physiology. Imaging techniques might also be used as a surrogate outcome measure in future thrombolytic trials. The context of a controlled study is the best environment to validate emerging imaging and treatment techniques.

The final section details reporting standards for complications and adverse outcomes; defines serious adverse events, adverse events, and unanticipated adverse events; and describes severity of complications and their relation to treatment groups. Recommendations are made regarding comparing treatment groups, randomization and blinding, intention-to-treat analysis, quality-of-life analysis, and efficacy analysis.

This document concludes with an analysis of general costs associated with therapy, a discussion regarding entry criteria, outcome measures, and the variability of assessment of the different stroke scales currently used in the literature is also featured.

Conclusion— In summary, this article serves to provide a more uniform set of criteria for clinical trials and reporting outcomes used in designing stroke trials involving intra-arterial thrombolytic agents, either alone or in combination with other therapies. It is anticipated that by having a more uniform set of reporting standards, more meaningful analysis of the data and the literature will be able to be achieved.

Key Words: stroke, ischemic • thrombolysis • trial design

	Introduction

Top Abstract Introduction IV Trials in Acute... IA Trials in Acute... Limitations of IA Thrombolysis Predictors for Treatment Success Pretreatment Evaluation Defining Patient Selection Defining Outcome of Therapy Treatment Description Posttreatment Evaluation:... Complications and Adverse Events Comparison Between Treatment... Costs Conclusions References

 
Stroke is the third leading cause of death in the United States, Canada, Europe, and Japan. According to the American Heart Association and the American Stroke Association, there are now >700 000 new strokes that occur each year, resulting in >200 000 deaths per year in the United States alone.¹ Ischemic stroke accounts for 80% and hemorrhagic stroke accounts for 20% of this total. Stroke is the leading cause of adult disability in North America and the No. 1 cause for inpatient Medicare reimbursement for long-term adult care.^2,3 The National Institutes of Health (NIH) estimates that stroke costs now exceed $45 billion in US healthcare dollars per year. At this time, the only therapy demonstrated to improve outcomes from acute ischemic stroke is thrombolysis of the clot responsible for the ischemic event. Intravenous (IV) thrombolysis has been studied in multiple trials, as described subsequently. Intra-arterial (IA) thrombolysis has been studied in 2 randomized trials and multiple case series. The purpose of this document is to provide guidance on study design for trials of IA cerebral thrombolysis of acute ischemic stroke. Included in this guidance document are reporting standards for such trials. Although many of the definitions and standards might be helpful in clinical practice, the study design recommendations are mainly intended for research trials. Many of the study design recommendations might also not be applicable for publications involving case series rather than controlled trials. However, the reporting standards should be applicable to all publications.

	IV Trials in Acute Ischemic Stroke

Top Abstract Introduction IV Trials in Acute... IA Trials in Acute... Limitations of IA Thrombolysis Predictors for Treatment Success Pretreatment Evaluation Defining Patient Selection Defining Outcome of Therapy Treatment Description Posttreatment Evaluation:... Complications and Adverse Events Comparison Between Treatment... Costs Conclusions References

 
In the late 1980s and 1990s, 8 large, well-designed, multicenter trials of IV thrombolysis were reported. Three initial trials of IV streptokinase (SK) in acute ischemic stroke, within 4 to 6 hours from symptom onset, were stopped by their safety committees owing to a high rate of acute mortality and intracranial hemorrhage (ICH) in the treatment arm. These included the Multicenter Acute Stroke Trial-Europe (MAST-E),⁴ the Australian Streptokinase Trial (ASK),⁵ and the Multicenter Acute Stroke Trial-Italy (MAST-I).⁶ Five other major studies evaluated the use of IV recombinant tissue plasminogen activator (rtPA) for acute stroke, from 3 to 6 hours from symptom onset, and included 2 trials sponsored by the National Institute of Neurological Disorders and Stroke (NINDS),⁷ the European Cooperative Acute Stroke Study (ECASS-I⁸ and ECASS-II),⁹ and the Alteplase Thrombolysis for Acute Noninterventional Therapy (ATLANTIS).¹⁰

The only 2 successful IV thrombolysis stroke trials to date have been the NINDS rtPA trials.⁷ These were placebo-controlled, randomized trials, performed as 2 separate studies, with 0.9 mg/kg rtPA given IV within 3 hours of symptom onset. Enrollment was completed in the first trial before starting the second trial with nearly identical design, including the same outcome measures but different prespecified end points. For practical reasons, the 2 trials were designated as Part A and Part B and analyzed as a meta-analysis. Nonetheless, they had extremely similar results, each successful in indicating superior outcomes for IV rtPA. The NINDS trial was a nonangiographic study involving 624 patients and included all ischemic stroke subtypes (small-vessel, large-vessel, and cardioembolic) of any degree or severity (median NIH Stroke Scale [NIHSS] score 14). Blood pressure was carefully controlled and maintained at <185/110 mm Hg. Patients meeting all entry criteria were treated with IV rtPA within 3 hours and were at least 30% more likely to have minimal or no disability at 3 months than were patients who received placebo. Benefit was seen in all stroke subtypes. Post hoc analysis failed to reveal any subgroup that did not benefit from IV thrombolysis, although the magnitude of improvement was less in elderly patients with severe strokes. However, there was a 10-fold increased risk of symptomatic brain hemorrhage within the first 36 hours: 6.4% in the rtPA group versus 0.6% in the placebo group (P<0.001). Although half of the brain hemorrhages were fatal, there were no differences in overall mortality between the 2 treatment groups. Early brain computed tomography (CT) changes of infarction and more severe neurologic deficit (NIHSS>20) were associated with an increased risk of symptomatic brain hemorrhage. The favorable outcome of these 2 NINDS-sponsored trials constituted the basis for US Food and Drug Administration (FDA) approval of IV rtPA in June 1996 for patients presenting with an ischemic stroke within 3 hours of symptom onset. To date, these 2 NINDS rtPA stroke trials have been the only positive, multicenter, randomized, controlled acute stroke studies in which IV lytics have been used.

ECASS-I⁸ was a multicenter, controlled study of IV rtPA for acute ischemic stroke conducted at 75 centers in Europe. A slightly higher dose of rtPA, 1.1 mg/kg, and a longer time window, 6 hours, were used in ECASS-I. Other important differences between ECASS and the NINDS trial were the inclusion of patients with moderate to severe hemispheric strokes only and the exclusion of patients with early CT changes involving more than one third of the territory of the middle cerebral artery (MCA). Despite careful selection of patients by trial examiners, 17.4% of the patients enrolled (109 of 620) had serious protocol violations, and more than half had major early infarct signs on CT. There was no benefit in the "intention-to-treat" (ITT) population in primary outcome by use of measures of functional outcome (the Barthel Index [BI] and modified Rankin Scale score [mRSS]) at 3 months. There was a "trend for improvement" in patients treated with rtPA within 3 hours of stroke onset. All of the secondary end points were significantly improved in the rtPA group, including speed of neurologic recovery, in-hospital stay, and the combined mRSS-BI score. Parenchymal hematomas occurred in 19.8% of the rtPA group and in 6.5% of the placebo group (P<0.001). Mortality was also significantly worse at 90 days in the rtPA-treated patients (P=0.04). Outcome was especially poor in the protocol violators who received rtPA, in which 42% died. Although the ITT analysis was negative, ECASS defined a "target" population by excluding protocol violators. In the target population, there was a significant improvement in the 90-day RSS (a primary outcome measure) in rtPA-treated patients (P=0.035). Also, when the NINDS statistical methodology was applied to the ECASS ITT dataset, there was a significant increase in favorable outcome in rtPA-treated patients, similar in degree to that seen in the NINDS study.

The ECASS II⁹ trial was a second attempt to demonstrate the efficacy of the 6-hour time window of IV rtPA for acute ischemic stroke. ECASS-II reduced the IV rtPA dose to 0.9 mg/kg (as in the NINDS trial). All participating sites underwent CT training sessions, and the rate of CT violations was reduced by 50%. Blood pressure parameters were also more tightly controlled. The parenchymal hemorrhage rate in patients receiving IV rtPA (11.8%) was lower than that seen in ECASS-I, although it was still 4 times more common than in placebo-treated patients (3.1%). Protocol violations declined to 9%. Unfortunately, there was no significant difference in the percentage of patients achieving the primary outcome measure of minimal or no disability (mRSS of 0 to 1 at 90 days) between the IV rtPA patients (40.3%) and the control group (36.6%, P=0.277). This reflected the better-than-expected outcome in the placebo arm. A post hoc analysis did demonstrate a significant (8%) increase in patients with slight or no disability (mRSS of 2) among treated patients.

The ATLANTIS trial^10,11 was also designed to demonstrate that the time window for IV rtPA administration could be extended to 3 to 6 hours from symptom onset. However, because of a high rate of brain hemorrhages, the treatment window was eventually decreased to 3 to 5 hours from stroke onset. The inclusion and exclusion criteria for this multicenter, randomized trial were very similar to those of the NINDS trial, except for the time window and exclusion of patients with early infarct signs of more than one third of the MCA territory, which was learned from the ECASS-I study. This trial was halted prematurely, after 63% of the planned patients had been enrolled, when an interim analysis indicated that treatment was unlikely to prove beneficial. Symptomatic intracerebral hemorrhage (ICH) occurred in 7% of IV rtPA patients versus 1.1% in the placebo group (P<0.001). However, in the 61 patients who were randomized to IV rtPA or placebo within 3 hours of symptom onset, there was a statistically significant 35% absolute increase in the number of patients with an NIHSS score of 1. Symptomatic (and fatal) ICHs occurred in 13% of treated patients and 0% of placebo patients. The death rate was more than 3 times higher in treated patients (17% versus 5%), but with the small numbers of patients involved, this difference did not reach statistical significance.¹²

	IA Trials in Acute Ischemic Stroke

Top Abstract Introduction IV Trials in Acute... IA Trials in Acute... Limitations of IA Thrombolysis Predictors for Treatment Success Pretreatment Evaluation Defining Patient Selection Defining Outcome of Therapy Treatment Description Posttreatment Evaluation:... Complications and Adverse Events Comparison Between Treatment... Costs Conclusions References

 
IA thrombolysis provides an alternative to IV thrombolysis in selected patients with acute ischemic stroke. Recent advances in the field of neurointerventional radiology, with the development of extremely soft, compliant microcatheters and steerable microguidewires; high-resolution fluoroscopy and digital imaging; and nonionic contrast agents, have made it feasible and safe to access the major intracranial blood vessels around the circle of Willis from a percutaneous transfemoral approach under local anesthesia. Rapid, local delivery of fibrinolytic agents or immediate access of thrombolytic devices is now feasible with these techniques and is performed at many major medical centers in selected patients with acute cerebral ischemia.

IA thrombolysis has been used most successfully in patients with acute MCA occlusion. There is evidence that the treatment window for IA thrombolysis with these techniques can be extended to at least 6 hours from stroke onset in patients with MCA occlusion. Other potential candidates for IA thrombolysis include patients with extracranial internal carotid artery (ICA) occlusion, intracranial carotid artery "T" occlusion, or basilar artery occlusion. Initial clinical and radiologic patient selection criteria and definition of outcome, as well as pretreatment and posttreatment evaluation, are mostly identical for IV and IA trials, although treatment description and complications differ significantly. Two randomized, multicenter, controlled trials of IA thrombolysis in acute MCA stroke have been reported so far, the Prolyse in Acute Cerebral Thromboembolism Trial (PROACT-I)¹³ and PROACT-II.¹⁴

Recanalization efficacy and safety of IA recombinant pro-urokinase (r-proUK) for MCA occlusion of <6 hours’ duration was demonstrated in the PROACT-I trial.¹³ The follow-up clinical efficacy trial, PROACT-II,¹⁴ was started in February 1996 and completed in August 1998. PROACT-II used an open design with blinded follow-up. Patients were screened with conventional angiography for occlusion of the MCA and had to have an NIHSS score between 4 and 30. The patients in PROACT-II had a very high baseline stroke severity with a median NIHSS of 17. Patients with early signs of an infarct in more than one third of the MCA territory on the baseline CT scan were excluded from the study. In PROACT-II, 180 patients were randomized 2:1 to receive either 9 mg r-proUK directly into an angiographically documented MCA occlusion plus low-dose IV heparin (2000-U bolus + 500 U/h x 4 hours) or low-dose IV heparin only. The primary outcome measure was the percentage of patients who achieved an mRSS 2 at 90 days, which signified slight or no neurologic disability. Secondary measures included the percentage of patients who had an NIHSS 1 at 90 days, angiographic recanalization, symptomatic ICH, and mortality. The median time from onset of symptoms to initiation of IA thrombolysis was 5.3 hours. In the r-proUK–treated group, there was a 15% absolute benefit in the number of patients who achieved an mRSS 2 at 90 days (P=0.043). Therefore, on average, 7 patients with an MCA occlusion would require IA r-proUK for 1 to benefit for the primary end point of achieving an mRSS 2 (mild or no disability). However, it is quite possible that others might also benefit in ways not captured by the primary end point (eg, improvement from mRSS 2 to 1 or from 4 to 3). The benefit was most noticeable in patients with a baseline NIHSS between 11 and 20. Recanalization rates were 66% after the 2-hour infusion for the treatment group versus 18% for the placebo group (P<0.001). Symptomatic brain hemorrhage occurred in 10% of the r-proUK group versus 2% in the control group. In PROACT-II, as in the NINDS trial, despite the higher early symptomatic brain hemorrhage rate, patients overall benefited from therapy, and there was no excess mortality (r-proUK 24%, control 27%) in the ITT analysis. The results of PROACT-II, though encouraging, were insufficient to secure FDA approval of r-proUK. The FDA has requested another larger trial of IA thrombolysis. Based on the available evidence, IA thrombolysis has been endorsed by several national organizations as an acceptable alternative stroke therapy in selected patients with acute ischemic stroke.

A third randomized trial was the Emergency Management of Stroke (EMS) Bridging Trial. This phase 1 pilot trial randomized patients either to treatment with reduced-dose IV alteplase, followed by arteriography, with IA infusion of up to 22 mg of alteplase if an appropriate arterial occlusive lesion was discovered (IV/IA), or placebo followed by IA lysis. In one third of cases, arteriography did not confirm the clot. Overall, of 15 patients with M1 or M2 occlusions, 66% good outcomes were achieved with a mean time to IA treatment of 4.2 hours. The IV/IA group had better recanalization than the IA group, but there was no difference in outcomes between the 2 treatment groups.¹⁵

	Limitations of IA Thrombolysis

Top Abstract Introduction IV Trials in Acute... IA Trials in Acute... Limitations of IA Thrombolysis Predictors for Treatment Success Pretreatment Evaluation Defining Patient Selection Defining Outcome of Therapy Treatment Description Posttreatment Evaluation:... Complications and Adverse Events Comparison Between Treatment... Costs Conclusions References

 
A major issue regarding access to IA thrombolysis to treat acute ischemic stroke is that it requires the ready availability of an endovascular interventionalist trained in IA thrombolysis and a stroke team. Such expertise is not currently available in most community hospitals across the United States and has usually been limited to large academic centers.^16,17 Another limitation of IA thrombolysis is the additional time required to begin treatment compared with IV thrombolysis. In PROACT-II, the average time from arrival at the hospital to the initiation of IA r-proUK was 3 hours. The time from completion of the CT scan to the start of lytic therapy was in the range of 30 to 90 minutes. However, even for IV tPA, the CT to needle time reportedly ranges from 52 to 68 minutes.^18,19 There are also concerns regarding the invasiveness of the technique and procedural risks not inherent to IV thrombolysis. However, serious procedural complications were uncommon in PROACT-I and II. Cerebral angiography in experienced centers is associated with a morbidity rate of 1.4% and a rate of permanent neurologic complications and death of 0.06% to 0.5%, respectively.^20–23

	Predictors for Treatment Success

Top Abstract Introduction IV Trials in Acute... IA Trials in Acute... Limitations of IA Thrombolysis Predictors for Treatment Success Pretreatment Evaluation Defining Patient Selection Defining Outcome of Therapy Treatment Description Posttreatment Evaluation:... Complications and Adverse Events Comparison Between Treatment... Costs Conclusions References

 
One of the most important predictors of clinical success is time to treatment, with early treatment of <3 hours’ demonstrating significant improvement in terms of 90-day clinical outcome and reduced cerebral hemorrhage.^{8–11,13,14,16,17} However, there are remaining problems in the assessment of patients who are eligible for thrombolysis. These include being able to predict whether a particular occlusion can be successfully revascularized, an individual patient’s prognosis and outcome after revascularization, and in particular, the development of ICH, with and without clinical deterioration. These predictors are important in selecting the appropriate target population for lytic therapy.

The immediate parameter for assessing treatment success is to evaluate the achieved flow restoration. However, it is not clear to presume that technical success, ie, recanalization, necessarily implies clinical success, ie, improved outcome. There is no simple correlation between recanalization and observed clinical benefit in ischemic stroke. From a pathophysiologic viewpoint, arterial occlusion is the cause of the stroke, and the primary mechanism of action of thrombolysis is clot lysis, resulting in recanalization and reestablishment of distal blood flow. However, although many case series have reported a "general relationship" between recanalization and good clinical outcome, this relation is known to be complex and dependent on several interactive variables. Ringelstein et al²⁴ reported that early recanalization of MCA occlusion within 8 hours of stroke onset had a favorable effect on infarct size and clinical outcome, but only in conjunction with good transcortical collateral blood flow. Von Kummer et al²⁵ reported that MCA partial or complete recanalization at <8 hours had no independent predictive value for good outcome and did not independently affect mortality. However, recanalization, even if delayed to 24 hours, was associated with improved clinical outcome in a subset of the patients studied, who were characterized by less hypodensity on baseline CT, baseline lower neurologic assessment, proximal site of occlusion, and good collateral flow. Similarly, a post hoc analysis of the PROACT-II data found that the relation between recanalization and clinical outcome was dependent on the precise site of arterial occlusion and on collateral arterial supply.²⁶ Improved outcome was associated with more distal occlusions and with better collateral flow demonstrated angiographically.

The appropriateness of using recanalization as a surrogate marker of outcome has had a correlation with recent diffusion and perfusion magnetic resonance (MR) studies. However, there is no direct correlation between recanalization and clinical outcome. Other factors in addition to recanalization, including baseline stroke severity, collateral circulation, lesion location, lesion volume, and time from stroke onset, are probably important in determining clinical outcome. Baird²⁷ identified a perfusion-diffusion mismatch on MR imaging (MRI) within 24 hours of stroke onset in 90% of patients with an arterial-occlusive lesion (ICA or MCA) on MR angiography (MRA). A mismatch between the infarct lesion size as shown by diffusion-weighted imaging (DWI) and that shown by perfusion imaging might mark potentially salvageable ischemic brain tissue.²⁸ Resolution of the mismatch, ie, a decrease in the size of the ischemic lesion, occurred in 75% of patients with arterial recanalization on follow-up MRA, compared with 36% of patients with persistent occlusion. MR perfusion-diffusion mismatch has been correlated with infarct growth and adverse clinical outcome.^28,29

The recanalization rates and clinical outcomes of thrombolysis vary with the site of arterial occlusion and techniques involved in treatment. In general, thrombus confined to the MCA beyond the striate arteries or its distal branches has a much better outcome than when the thrombus also involves the supraclinoid carotid artery segment and then extends into the proximal middle and anterior cerebral artery (T-lesion occlusions).²⁵ Patients with ischemic stroke of <6 hours’ duration have a wide variety of arterial-occlusion sites, and 20% have no visible occlusion, despite similar neurologic presentations.^14,30 In PROACT-II, 474 patients had a screening angiogram to enroll 180 eligible patients with proximal MCA occlusions.¹⁴ In the IV thrombolysis stroke trials, neither the sites of arterial occlusion nor the recanalization rates were known. Angiography performed during IA lysis permits documentation of both the site of arterial occlusion and recanalization rates and provides access for IA thrombolysis.

Recanalization rates with IA thrombolysis are superior to those for IV thrombolysis for major cerebrovascular occlusions. Recanalization rates for major cerebrovascular occlusions average 70% for IA thrombolysis compared with 34% for IV thrombolysis.³¹ The differences in recanalization rates are most apparent with large-vessel occlusions such as the ICA, which is the most difficult vessel for thrombolysis, followed by the carotid T segment and the proximal (M1) segment of the MCA.²⁵

There have been no randomized studies comparing recanalization rates and clinical outcomes between IV thrombolysis and IA thrombolysis. Limited data suggest, however, that IV rtPA might be relatively ineffective in patients with ICA or MCA occlusion. Data from the TTATTS (Thrombolytic Therapy of Acute Thrombotic/Thromboembolic Stroke) study indicate that the recanalization rate for large-vessel occlusion with 0.8 mg/kg or 1.0 mg/kg IV rtPA is no more than 30% effective (personal communication). Tomsick et al³² reported that IV rtPA given <3 hours from stroke onset was ineffective in patients with a baseline NIHSS score 10 and a hyperdense MCA sign (signifying MCA occlusion) on brain CT scans.

Prediction of hemorrhagic complications is another unresolved problem with thrombolytic treatment. Aggregate data indicate an 8.3% risk of symptomatic brain hemorrhage with IA thrombolysis in the carotid territory and a 6.5% risk in the vertebrobasilar territory.³³ There is no evidence that the rate of symptomatic brain hemorrhage is lower with IA thrombolysis than with IV thrombolysis, but direct comparisons are difficult. In an uncontrolled series, Gönner et al³⁴ reported a 4.7% rate of symptomatic brain hemorrhage in 42 patients treated with IA thrombolysis. This series differed from PROACT-II in that only 26 of the 42 patients received heparin; the remainder received aspirin. The higher rate of ICH causing neurologic deterioration with IA r-proUK in PROACT-II (10.2%)¹⁴ compared with IV rtPA in NINDS (6.4%),⁷ ATLANTIS (7.2%),¹⁰ and ECASS-II (8.8%)⁹ must be understood within the context of the greater baseline stroke severity, longer time to treatment, and 66% MCA recanalization rate in PROACT-II. However, although brain hemorrhage complicating thrombolysis for acute stroke likely reflects reperfusion of necrotic tissue, several series have found no direct relation between recanalization and hemorrhage risk.^35,36 The amount of ischemic damage is a key factor in the development of brain hemorrhage after thrombolysis-induced recanalization. Major, early CT changes and severity of the initial neurologic deficit, both indicators of the extent of ischemic damage, are some of the best predictors of the risk of hemorrhagic transformation.^8,36 The median baseline NIHSS score in ATLANTIS and ECASS-II was 11; in NINDS, 14; and in PROACT-II, 17. Greater baseline stroke severity was first associated with increased ICH risk in NINDS and ECASS-I. All symptomatic ICHs in PROACT-II occurred in patients with a baseline NIHSS score 11, and in NINDS, the rate of symptomatic brain hemorrhage in patients with an NIHSS >20 was 18%.

The dose of the thrombolytic agent,³⁷ blood pressure,^7,38,39 advanced age,³⁹ prior head trauma,⁴⁰ and blood glucose >200 mg/dL⁴¹ have been associated with hemorrhage after thrombolysis for both stroke and myocardial infarction Adjunctive antithrombotic therapy might also play a role during IA thrombolysis. Age was the most important risk factor in 1 of the largest series of thrombolysis-related ICHs in patients treated for coronary occlusions.⁴⁰ A relation between advanced age and hemorrhage was demonstrated in the NINDS³⁶ and ECASS trials.^8,9 Although there is no strict age cutoff for administering thrombolytics for stroke, physicians need to take age into account, especially in patients older than 75 years when determining the risk of angiography and IA thrombolysis.

Summary
Most lessons learned in previous studies have been immediately implemented in the protocol of a subsequent trial, eg, standardized baseline neurologic examination (the NIHSS score), exclusion of patients with extensive early CT signs, dose corrections for both thrombolytic agents and concomitant anticoagulation and antiplatelet medications, or training diagnostic radiologists and neurologists to accurately read screening CT brain scans. These should be standards for future thrombolytic trials and are mentioned in detail in the succeeding sections of this report.

Other exploratory analyses of thrombolytic studies create new hypotheses by recognizing parameters that might influence treatment success and outcome. For example, the precise clot location, even within the MCA territory, or the extent of tissue-preserving perfusion, eg, via collateral supply, is of importance. The interaction between the different clinical and radiologic parameters is complex, and their combined implication for treatment success and outcome is still the subject of ongoing investigations. To validate the importance of the different variables, they should be collected in upcoming trials and analyzed post hoc. These imaging-based criteria for the assessment of collateral flow might possibly override our current time-based selection rules for inclusion. For example, a patient with evidence of good collateral flow might still be a candidate for thrombolysis, even when presenting beyond the accepted time window of postsymptom onset, whereas patients without evidence of any collateral flow or tissue perfusion might be excluded from trials even when they present within the generally accepted time window. This document provides a framework of study design and reporting standards by which these new hypotheses might be evaluated.

	Pretreatment Evaluation

Top Abstract Introduction IV Trials in Acute... IA Trials in Acute... Limitations of IA Thrombolysis Predictors for Treatment Success Pretreatment Evaluation Defining Patient Selection Defining Outcome of Therapy Treatment Description Posttreatment Evaluation:... Complications and Adverse Events Comparison Between Treatment... Costs Conclusions References

 
To provide information necessary for defining patient selection, inclusion/exclusion criteria, and outcomes of therapy, it is necessary to describe the methods used to evaluate stroke patients before treatment. Such pretreatment evaluation includes general medical history and physical examination, neurologic history and examination, laboratory (blood) tests, and imaging.

General Medical Evaluation
A complete physical examination and medical history should be obtained at screening, including a detailed cardiovascular and cerebrovascular medical history. Blood pressure (3 successive measurements 10 minutes apart), pulse rate, and respiration rate should be obtained at screening. Any medication (including over-the-counter medicines such as aspirin, antacids, vitamins, mineral supplements and herbal preparations) that the subject has taken within 48 hours before enrollment or has received during the first 7 days (or hospital discharge) after enrollment should be recorded, along with the dates of administration, dose, and frequency.

A 12-lead ECG is recommended before the procedure, at 24 hours, and then at 48 to 72 hours after randomization when indicated. Samples for cardiac enzyme levels, to rule out acute myocardial ischemia/infarction, should be drawn as appropriate when any abnormal ECG changes are detected. In addition, cardiac echocardiography is recommended in all patients with a history of arrhythmias.

Laboratory Evaluation
Before angiography, blood specimens for the following laboratory studies should be obtained: (1) Hematology: hematocrit, hemoglobin, and platelet count; (2) Coagulation parameters: activated partial thromboplastin time (aPTT), prothrombin time, and international normalized ratio. In the event that the subject was receiving heparin therapy just before screening, an aPTT must be performed with a result 1.5 times the upper limit of normal before randomization; Clinical chemistry: Tests of liver function (aspartate aminotransferase, alanine aminotransferase, alkaline phosphatase) when indicated; kidney function (creatinine, blood urea nitrogen), and serum glucose.

Neurologic Evaluation
Examples of neurologic impairment scales include the NIHSS,⁴² Canadian Neurologic Scale,⁴³ and Scandinavian Stroke Scale.⁴⁴ The NIHSS is a 42-point scale that quantifies neurologic function in specific categories and provides a means to measure neurologic deficits. Higher scores indicate a more severe neurologic deficit. The NIHSS was developed by researchers at the NINDS specifically for use in clinical stroke trials. The NIHSS has been extensively used in clinical trials to quantitatively measure stroke outcomes and has been validated and standardized to reduce interobserver error. Neurologic assessments should use the NIHSS. The screening NIHSS assessment should be used for subject entry restriction of NIHSS score. Screening NIHSS should be performed by an examiner experienced in acute stroke treatment who has achieved certification in administering the NIHSS. The NIHSS score should be determined again just before diagnostic cerebral angiography. If the NIHSS score has significantly improved from baseline, such as a 4-point improvement, or has improved below or increased above the defined threshold, the subject might not be eligible for randomization. The NIHSS was designed primarily for anterior-circulation ischemic stroke evaluation. It might underestimate the severity of deficit for posterior-circulation strokes, because posterior-circulation symptoms, such as vertigo or difficulty in swallowing, are not included on the NIHSS evaluation. A standardized grading system for posterior-circulation strokes is not currently available.

Disability and handicap scales include such instruments as the Barthel Index (BI) of Activities of Daily Living,⁴⁵ the modified Rankin Scale (mRS),⁴⁶ and the Stroke Impact Scale.⁴⁷ There is a written examination to certify the examiner in administering the BI. There is no certification procedure for the mRS, which is the most commonly used global assessment of stroke outcome. All of these stroke outcome assessments are reliable, familiar to the stroke neurology community, and adaptable for use in patients with acute stroke and can be used to compare outcomes as end points when evaluating other published trials of thrombolytic therapy. The BI is used for functional evaluation to measure activities of daily living and has been used since 1955 as a simple index of independence by scoring the ability of patients with neuromuscular or musculoskeletal disorders and by evaluating progress in these patients. The BI scores patients in each of 10 listed activities, and a score of 10 or 15 is assigned for each activity that a patient can perform independently. Patients who score 100 on the BI are continent, can feed and dress themselves, get out of bed, walk >1 block, and perform activities of daily living. The mRS is used to measure overall functional disability and handicap after a stroke. The original Rankin scale was modified in 1988 by Warlow et al for the UK-TIA study to accommodate language disorders and cognitive deficits. A score of <2 on the mRS is considered a favorable outcome with minimal or no disability. At screening, a historical BI and a historical mRSS should be obtained from the subject or subject’s caretaker. These determinations should reflect the functional and disability status of the subject before the stroke. These historical determinations do not have to be obtained before enrollment of the subject.

Imaging Evaluation
Imaging techniques such as CT, MRI, and angiography have an important role in screening and monitoring of patients treated with IA thrombolysis, ie, before, during, and after intervention. Imaging evaluates gross anatomic abnormalities, such as the presence of ICH or infarction, as well as vascular perfusion and collateral flow. To reliably evaluate inclusion and exclusion criteria or to assess treatment success, it is important that the imaging examinations be performed according to state-of-the-art technology and are interpreted according to standardized rules.

Angiography
The gold standard for the demonstration of a vascular occlusion in IA thrombolysis trials is angiography. The confirmation of a vascular occlusion seems to be quite important, knowing that up to 20% of patients with suspected cerebral ischemia will have a negative angiogram (possibly due to early spontaneous clot lysis or microvessel occlusion) and another 10% to 20% have other angiographic exclusion criteria.^14,48 Also, patients with an acute ischemic stroke might have a variety of arterial occlusion sites despite similar clinical presentations.³⁰

A complete diagnostic cerebral angiogram of the affected territory should be obtained at baseline. For purposes of investigational trial design, a three or 4 vessel pre-intervention diagnostic cerebral angiogram, including both internal carotid arteries and the dominant vertebral artery, which includes the late venous phase is recommended before thrombolysis to evaluate concomitant pathologies and anatomic variations and to assess collateral flow from all possible sources. In addition, if significant aortic arch disease is suspected, aortic arch angiography should also be performed. Other important parameters to evaluate from a baseline angiogram would be the presence or absence of shunting (appearance of a vein during arterial phase before venous phase) and the presence of a vascular blush, both of which might indicate an increased risk for ICH.^49,50

Location of Occlusion
The proximal extent of occlusion is correlated with technical success of lysis, neurologic recovery, and risk of ICH.^25,26,50 The location of occlusion can be reported as described in Table 1.

View this table: [in this window] [in a new window]   TABLE 1. Location of Occlusion

Perfusion
The Thrombolysis in Myocardial Infarction (TIMI)⁵¹ definition was originally adapted and used to describe flow in the coronary arteries from 0 to 3. Its use has been extended to angiographic cerebral blood flow.¹⁴ Recently, Thrombolysis in Brain Ischemia (TIBI) definitions have become available from transcranial Doppler data.⁵² To emphasize the use of a standard grading system for a thrombolytic trial specific to the intracranial cerebral circulation, we propose a Thrombolysis in Cerebral Infarction (TICI) grading system (Table 2).

View this table: [in this window] [in a new window]   TABLE 2. Thrombolysis in Cerebral Infarction (TICI) Perfusion Categories

Collateral Flow
In the absence of direct perfusion, tissue is preserved by the presence of collaterals. Consequently, it is expected that infarct development and the result of thrombolytic drug treatment depend on the presence of collaterals. It has been stated that the presence and size of the ischemic penumbra are influenced by collateral circulation.⁵³ In PROACT-II, for example, the presence of collaterals seemed to have a major effect on the CT appearance of an infarct, as well as on the clinical presentation. Thus, the more collaterals, the smaller was an infarct on CT and the smaller was the stroke scale score. The effect on outcome, however, seemed to depend on treatment as well: in the presence of collaterals, patients treated with IA thrombolysis had a better outcome compared with the control group. In the absence of collaterals, r-proUK did not improve 90-day outcome compared with control. From both outcome and CT data, it was concluded that patients without collaterals might not benefit from drug treatment, whereas patients with collaterals do indeed improve after r-proUK treatment, compared with controls, and that CT findings corroborate the 90-day-outcome result.²⁶ Consequently, it might be important to quantify collateral flow and use this measure in future trials as an inclusion criterion.

The effect of collaterals might also be influenced by location of the occlusion. For example, a proximal MCA clot might occlude the lenticulostriate arteries that supply the basal ganglia and internal capsule. These vessels are not collateralized from the cortex. Therefore, excellent transcortical collaterals might allow cortex to be salvaged, but permanent hemiplegia might still result from infarction of the internal capsule.

In the literature, there is insufficient information on grading systems for collateral flow. The "gold standard" for the assessment of collateral flow is angiography. An estimation of collateral flow was performed in the PROACT-II trial by evaluating the angiographic presence or absence of flow to the ischemic site and to the occlusion site.²⁶ A proper grading of collateral circulation is essential for future clinical trials. We propose that the function of the collateral circulation be evaluated with the grading system shown in Table 3.

View this table: [in this window] [in a new window]   TABLE 3. Collateral Flow Grading System: Angiographic

Time of collateral filling is somewhat subjective; however, it might be determined by counting the number of frames from contrast-agent filling of the petrous carotid artery (from the anterior circulation) and the proximal basilar artery (in the posterior circulation) to complete collateral filling (provided that the number of frames per second is known, so that the time for collateral filling can be calculated.) This is then compared with time to normal filling of the nonoccluded hemisphere in the parenchymal phase of the angiogram. Slow collateral flow is defined, arbitrarily, as filling that is >2 seconds slower than the contralateral side. Rapid collateral flow is defined as filling that is within 2 seconds of the contralateral side. It is essential that the angiogram include both the arterial and venous phases of the injection to evaluate the collateral pathways.

Computed Tomography
The noncontrast CT scan still is regarded as the most important diagnostic tool in the assessment of patients with a suspected acute stroke⁵⁴ to exclude hemorrhage and demonstrate early infarct signs. A method of quantifying the volume of early infarct signs from the screening CT scan has been suggested.⁵⁵ The baseline (screening) CT brain scan should consist of a nonenhanced CT scan with contiguous, discrete (nonhelical) images from the vertex of the calvarium through the foramen magnum. Sections that are 5 mm thick are preferred, but scans should not be thicker than 10 mm. The scan plane should be parallel to the canthal meatal line. Ideally, the screening CT should be performed no longer than 1 hour before initiating thrombolytic therapy. Otherwise, there might be ischemic or hemorrhagic changes that have developed from the time of initial head CT that would have excluded the patient from therapy

ECASS-I has shown a high rate of protocol violations regarding the baseline CT inclusion criteria.⁸ Thus, as implemented in ECASS -II,⁹ it is recommended that all sites should undergo CT training sessions, so that the reader can detect early CT changes with a high degree of sensitivity and reliability.

Recently, CT scans have been expanded to obtain functional information on the state of the tissue and the location of a vascular clot. As such, CT angiography and a dynamic CT perfusion scan might become part of the screening CT protocol. A CT angiogram, if it can be expeditiously obtained, is recommended to cover the entire cerebrovascular axis, including the anterior (carotid) and posterior (vertebrobasilar) circulation, as well as the extracerebral carotid arteries.⁵⁶ A helical scan is required for CT angiography; however, multidetector scanners are preferred. Scanning can be performed from the vertex to the aortic arch during the injection of 100 mL of contrast medium.⁵⁶ Vascular opacification can be observed on the cross-sectional images and displayed in 2D and 3D reformations. In the context of controlled, clinical trials of IA thrombolysis, it is possible to validate CT perfusion and CT angiography for the assessment of the tissue and vasculature, in particular, to evaluate and grade the collateral flow. Such a noninvasive, validated screening technique might then in future trials be used to reduce the number of angiograms.

A dynamic perfusion CT scan acquires data at a single location continuously during the transit of a contrast-agent bolus (40 mL) through the tissue.^57,58 Reviewing or postprocessing the acquired images allows the assessment of tissue perfusion by virtue of different parameters describing the contrast-agent transit curve. The CT perfusion technique was originally described in the early 1980s^59–61 but failed to gain general application. However, newer techniques, in particular, faster scanner generations, and the availability of thrombolytic agents revived interest in CT perfusion imaging.^{57,58,62–66} The term CT perfusion was initially adopted for the assessment of "perfused blood volume" with contrast-enhanced CT.⁶² This technique utilizes subtraction images of nonenhanced and contrast-enhanced images of the whole brain, which then indicate the perfused areas of cerebral tissue. After contrast enhancement, Hounsfield units will change in a reference blood vessel, and this allows interpretation of subtraction images, such as maps of fractional, or perfused, blood volume, and therefore, the delineation of an acute infarct might be reflected as a reduction in regional cerebral blood volume (rCBV). However, even in an acute ischemic stroke, CBV can be either increased, normal, or decreased, depending on the severity of hypoperfusion and collateral flow.⁶⁷

Thus, the CT perfusion technique has recently been expanded to assess the whole dynamics of a contrast-agent transit curve (dynamic CT perfusion). Assuming an intact blood-brain-barrier, the contrast agent yields a time-varying change in density or signal intensity. A variety of perfusion indices extracted from the analysis of the contrast-agent transit curve have been proposed based on empirical measures or slightly more sophisticated hemodynamic modeling (eg, arterial deconvolution). Most commonly used are the maximal-density/signal-intensity change (peak), time-to-peak density/signal intensity (TTP), the full width of the contrast-transit curve at half height, or mean transit time (MTT). Integrating the area under the concentration-time curve yields relative estimates of the rCBV, and the ratio of the rCBV and the MTT yields a relative estimate of regional cerebral blood flow (rCBF). Analysis software is commercially available, and algorithms have been developed to actually quantify flow in cerebral tissue.⁵⁸ In acute ischemia, dynamic CT perfusion has been shown to be a reliable alternative to MR perfusion.^{63,66,68–70}

The presence of collateral flow can be estimated on perfusion imaging by both CT and MR. A typical pattern of collateral flow consists of a lower peak, delayed TTP, increased MTT, decreased flow, and normal or elevated CBF. A typical pattern of absent collateral flow would be characterized by a decreased CBV, which has been shown to indicate a poor prognosis. Moreover, the combination of increased time to enhancement and increased rCBV indicates good collateral flow.

Magnetic Resonance Imaging
An MR brain study at baseline might be performed in addition to or instead of a CT study. However, the same study (CT or MR) should be performed at all scheduled time points. An MRI protocol is suggested to include a T2-weighted or fluid-attenuated inversion recovery sequence, a diffusion-weighted sequence, and a perfusion-weighted sequence. A 3D MRA of the circle of Willis can be performed without (eg, 3D time of flight) or with contrast agent. If a contrast-enhanced gadolinium MRA is performed, care should be taken to acquire the data in the arterial phase of the bolus to prevent venous opacification. MRA demonstrates patency or occlusion of cerebral arteries.

MRI has a high sensitivity and specificity for the diagnosis of ischemic stroke in the first several hours after symptom onset, identifies arterial occlusions, and noninvasively characterizes ischemic pathology.⁷¹ Case series have demonstrated and characterized the early detection of intraparenchymal hemorrhage^72–74 and subarachnoid hemorrhage^75–77 by MRI. Echo planar images, used for diffusion and, in particular, perfusion MRI, are inherently sensitive to the susceptibility changes caused by intraparenchymal blood products.^72–74 Consequently, MRI has replaced CT to rule out acute hemorrhage in some centers.

Much attention in recent years has been focused on using MRI for perfusion imaging.^29,78–83 On perfusion-weighted imaging the volume of ischemic brain is detected. Contrast-agent transit is observed by using dynamic susceptibility contrast-enhanced T2/T2*-weighted MRI. The same indices used for CT perfusion are used for MR perfusion.

A strong argument for MR perfusion imaging in acute ischemia is the possibility to combine the perfusion information with DWI, a technique that highlights the cytotoxic edema in the core of the infarcted brain. DWI allows detection of cerebral ischemia within minutes of onset, and the temporal evolution of diffusion characteristics enables differentiation of acute from chronic stroke.⁸⁴ On DWI, the volume of the infarcted lesion appears hyperintense relative to the surrounding tissue. When the region of hyperintensity on DWI is surrounded by a larger perfusion defect, this pattern is called a perfusion/diffusion mismatch.⁸² This mismatch indicates the presence of the ischemic penumbra, ie, ischemic tissue at risk that might go on to infarction without revascularization and that is potentially salvageable by flow restoration. MRI thus has tremendous potential for helping direct the treatment of acute ischemic stroke.

However, there are some practical limitations of MRI. Controversy still exists over the pathophysiology of underlying changes in diffusion^85,86 and the reversibility of changes after reperfusion.^87,88 It has been advocated to offer studies that take only 10 to 20 minutes to perform, but in reality, an MR examination in an acute stroke patient takes much longer, and sequences are often distorted by motion artifacts and have to be repeated. Patient monitoring is difficult, and the setup is time-consuming. Furthermore, in many institutions, MRI might not be available in an emergency situation.

In experienced centers, a 15-minute MRI session might help select subjects for IA thrombolysis based on a positive diagnosis of an occlusion by MRA or a perfusion defect that exceeds the region of acute injury demonstrated by DWI. Subjects without appropriate pathology on MR images are spared the risk of an unnecessary arteriogram. The combination of DWI and PWI, if it can be obtained in a reasonable time frame, is advocated as a tool in stroke clinical trials and also for assessing the drug effect on ischemic pathology.

Summary
Preprocedure neurologic assessment by a clinician certified in all of the applicable neurologic tests is required at specified time points. The examinations should include the NIHSS, BI, and mRSS assessments. Baseline laboratory studies that should be performed include a complete blood count with platelets, coagulation studies, liver function tests if indicated, and routine electrolyte blood studies.

Baseline imaging studies should include a noncontrast CT head scan and then possibly a CT angiogram, MRI/MRA, perfusion/diffusion imaging studies, and diagnostic angiography as specified in the clinical study protocol. Catheter angiography should include exact information on occlusion location and on collateral flow, which should be evaluated according to a uniform grading system.

Recommendations for reporting include age and gender of the patients enrolled; neurologic deficit baseline NIHSS; functional status by the mRSS and baseline BI; angiographic location, of the clot; baseline perfusion and collateral grade; and CT or MR assessment of baseline hemorrhage and edema. Assessment of perfusion (CT or MR) and diffusion (MR) is optional.

	Defining Patient Selection

Top Abstract Introduction IV Trials in Acute... IA Trials in Acute... Limitations of IA Thrombolysis Predictors for Treatment Success Pretreatment Evaluation Defining Patient Selection Defining Outcome of Therapy Treatment Description Posttreatment Evaluation:... Complications and Adverse Events Comparison Between Treatment... Costs Conclusions References

 
The selection of the study population should be based on clinical presentation (signs and symptoms of acute stroke presenting in the appropriate time window), CT, MRI, and/or angiography. The entry criteria (ie, inclusion and exclusion criteria) determine which patients will be enrolled into a clinical trial. They also influence patient baseline characteristics. Consequently, the entry criteria influence both clinical outcome and the likelihood of finding a statistically significant effect between treatment arms.⁸⁹ The entry criteria used in the previous thrombolytic trials were, in general, based on the same assessment scales, eg, NIHSS score, patient age, noncontrast CT scan, etc. Depending on the drug or device under study, a trial might be designed to avoid enrolling a large population of patients expected to have a very good or very bad outcome regardless of therapy, such as those with mild or very severe strokes.⁸⁹ When choosing the specific thresholds for inclusion and exclusion in future trials, it should be recognized that enrolling patients with a relatively high or low chance of having a good outcome might influence the ability of a study to demonstrate a treatment-related benefit,⁸⁹ and it might be useful to stratify patients according to variables that predict outcome, such as the NIHSS score. The inclusion and exclusion criteria listed in the following sections have proven useful in several reported major trials of MCA strokes but might not be completely applicable to future trials. For example, recent surgery is listed as a contraindication from previous trials, although a case series showed acceptable risks of cerebral thrombolysis in patients with recent major surgery, including cardiac surgery.⁹⁰