Acute Stroke Intervention

Acute ischemic stroke is a common cause of death and disability. To date, the only therapy proven to be effective is to restore blood flow by pharmacological fibrinolysis within the first few hours of stroke onset. The margin of benefit in the clinical trials of fibrinolysis has been small, however. This is due in part to the risk of intracranial hemorrhage after fibrinolytic therapy. Furthermore, many patients with acute ischemic stroke are not candidates for this treatment because of late presentation. Consequently, better and safer methods for the rapid dissolution of occlusive thrombus and better methods of patient selection based on advanced imaging have become two complementary areas of very active investigation. In this article, we will review the pathophysiology of acute ischemic stroke, the current proven therapies—intravenous (IV) tissue plasminogen activator (t-PA) and intra-arterial pro-urokinase (pro-UK)—ongoing research involving physiological imaging for patient triage, and new endovascular methods of revascularization


PATHOPHYSIOLOGY
Stroke is clinically defined as the sudden onset of an irreversible neurological deficit. These events are either hemorrhagic, bleeding into the substance of the brain or within the confines of the skull into the spaces around the brain, or ischemic, inadequate blood flow to support the metabolic needs of the brain. This distinction may be difficult to make on the basis of a clinical examination and rests almost entirely on a computed tomography (CT) examination. Most strokes are ischemic and most are due to embolic material arising in the heart or from the large arteries supplying the brain. 1,2 This material may consist of platelets, red cells, atheromatous debris, or a mixture of these elements. Emboli may also cause temporary neurological deficits, such as transient ischemic attacks, that resolve completely.
When an embolus breaks free from its source, it is carried downstream until it lodges in a smaller vessel. This often occurs at branch points of the intracerebral arterial tree. Whether permanent neurological injury (stroke) will occur after embolic occlusion depends on three interrelated factors: the duration and the degree of the reduction in blood flow 3 and the vulnerability of the affected cells to ischemic injury. 4 For any given level of reduced blood flow, the longer the duration, the greater the likelihood of permanent injury. Similarly, for any given duration of ischemia, greater degrees of flow reduction will increase the risk of stroke. The vulnerability of particular neurons to ischemia is variable: CA1 hippocampal neurons may die after 3 minutes of severe ischemia, whereas neurons in other areas of the brain subject to the same degree of ischemia may show no signs of infarction after ischemic periods of up to 15 minutes. 4 The actual physiological state of the brain and its vasculature in patients presenting with acute stroke may be variable and may change over time. The degree and duration of ischemia in an arterial territory beyond an occlusive or partially occlusive embolus may vary over different regions. This point relates to the concept of an ischemic core and penumbra. In the complete absence of blood flow, permanent injury or cell death may occur as soon as 3 minutes. All patients presenting with acute ischemic stroke, and many presenting with a transient ischemic attack, will have some dead brain cells within the core of the ischemic area. The real hope is to salvage other ischemic cells that are still viable. These viable but ischemic cells are referred to as the ischemic penumbra. The degree of ischemia may also vary in time. Many patients presenting with acute ischemic stroke have no identifiable occlusive lesion. It is now recognized that in the minutes, hours, and days after stroke, many embolic occlusions resolve spontaneously, without intervention, while some remain unchanged. [5][6][7] In one large multicenter trial of fibrinolysis, nearly one third of the patients with acute ischemic stroke undergoing emergent angiography before randomization had no evidence of arterial occlusion. 8 The response of the tissue to reperfusion, whether by natural fibrinolysis or by therapeutic intervention, may vary. First, the ischemic tissue may not recover. Some cells may be dead, others may initiate a sequence of intracellular events resulting in delayed or programmed cell death (apoptosis), and others yet may survive but not recover normal neurological functions. Clinical trials of agents aimed at preventing secondary injury or apoptosis have been universally and spectacularly unsuccessful, but the possibility remains that they may be useful in subgroups of patients with ischemic stroke. Another consequence may be hemorrhage. The blood-brain barrier, normally impermeable to red blood cells and highmolecular-weight proteins, can become disrupted by severe ischemia. This may lead to hemorrhagic transformation of an ischemic infarct. Hemorrhagic transformation may be asymptomatic but may also cause lifethreatening mass effect. This outcome is more common, and especially worrisome, in patients treated with fibrinolytic agents. Severe brain edema and swelling may be related to the amount of dead tissue, blood-brain barrier disruption, timing of reperfusion, and possibly fibrinolytic therapy.

CURRENT TREATMENT: INTRAVENOUS FIBRINOLYSIS
Controlled clinical trials have proved that IV fibrinolysis is an effective treatment for patients presenting with symptoms of acute ischemic stroke, despite the complicated physiological and anatomic factors discussed earlier. 9, 10 The magnitude of benefit in these trials is marginal, however. A major factor limiting the benefit of fibrinolysis is the increased risk for symptomatic intracerebral hemorrhage. Furthermore, the number of patients qualifying for treatment is low. The therapy is limited to patients presenting within 3 hours of onset. Some have estimated that less than 2% of patients with acute ischemic stroke are treated with fibrinolytic agents. 11 In this section, we will discuss these trials in detail, because of their impact and implications for future applications of imaging and intervention.
For 25 years thrombolytic agents, more properly called fibrinolytic agents, have been investigated as a treatment to improve clinical outcome after acute thromboembolic stroke. The basic principle of fibrinolysis is that a drug is delivered to increase the rate of lysis of an obstructive cerebral arterial clot, leading to faster restoration of normal cerebral blood flow to the ischemic tissue. Fibrinolytic agents, including streptokinase, urokinase, pro-UK, and t-PA, stimulate conversion of plasminogen to plasmin. Plasmin then converts fibrin, a substance found in acute clots, to fibrin split products, resulting in clot dissolution. Fibrinolytic drugs, dependent on this mechanism to restore flow, are more effective in dissolving fresh, fibrin-rich clots than chronic, organized, fibrin-poor clots.
The European Cooperative Acute Stroke Study (ECASS I) was conducted using a total IV t-PA dose of 1.1 mg/kg delivered within the first 6 hours of acute stroke symptoms to patients who had CT scans showing no hemorrhage. 12 The study showed a slight benefit in survivors and improved neurological outcome at 90 days in about 12% more of the patients treated with IV t-PA compared with controls, but it also showed a significantly higher rate of cerebral hematoma formation in the t-PA-treated patients (20%) compared with controls (7%) and a higher mortality in t-PA-treated patients (15%) than in controls (12%). No benefit was found overall, because of the increased complications in the treatment group.
The National Institute of Neurological Disease and Stroke (NINDS) trial of IV t-PA for acute stroke was conducted with a slightly lower dose, 0.9 mg/kg, than what was used in the ECASS trial, and the time limit for symptom duration prior to treatment was reduced to 3 hours from 6. 9 The NINDS trial showed a slight but significant clinical benefit at 90 days, 13% more t-PA-treated patients with a "good" modified Rankin score of 0-1 compared with controls. There were six times (6% versus 1%) the number of symptomatic cerebral hemorrhagic transformations in the t-PA-treated group than in controls, but that did not increase mortality in the t-PA-treated group. No benefit was seen in t-PAtreated patients at 24 hours.
A subsequent ECASS II study conducted with a 3-hour time limit of symptom duration prior to IV t-PA treatment supported the use of IV t-PA for acute nonhemorrhagic stroke within the shorter time frame, but only when patients without normal CT scans at the time of treatment were excluded from the analysis and only after a good neurological outcome was defined with an expanded modified Rankin scale of 0-2. 13 A subsequent NINDS trial of IV t-PA for acute thromboembolic stroke when the drug was given in the 3-to 5-hour time frame after symptom onset did not show a benefit for treatment. 14 In summary, clinical trials have shown that IV fibrinolytic therapy is marginally effective and only within a 3-hour time window. Nevertheless, it is currently the standard of care for selected patients presenting within 3 hours and with no CT evidence of hemorrhage.

CURRENT TREATMENT: INTRA-ARTERIAL FIBRINOLYSIS
The advantages of delivering the drug locally in the clotted artery in the brain through a small catheter are that a higher concentration of the lytic agent will reach the clot, increasing the effectiveness of clot lysis, and that a lower total dose of the lytic agent can be delivered, decreasing the risk of bleeding elsewhere in the body from systemic effects. To inject the lytic agent directly at the site of the clot in the cerebral artery obviously requires more time and expertise than delivering the agent into a peripheral vein does, but the added effort appears to be worthwhile.
To date, one controlled trial, the Prolyse for Acute Cerebral Thromboembolism (Pro-ACT II) trial, has shown a benefit for intra-arterial thrombolysis. 8 In this study, in which there was angiographic confirmation of an obstructive clot in the proximal middle cerebral artery, the fibrinolytic agent pro-UK was infused at the clot surface through a microcatheter for a fixed time and at predetermined doses. Mechanical disruption of the clot was not allowed. The study demonstrated that the extent of clot lysis depended on the doses of the fibrinolytic agent and heparin administered, with about 67% of the clots partially or completely dissolved during the procedures in the treated groups compared with about 10% in the control groups. The clinical benefit demonstrated for intra-arterial fibrinolysis was that more treated patients (40%) than control subjects (25%) were in good neurological condition 90 days after treatment, as defined by falling into the "0" and "1" categories on the modified Rankin scale. The rate of symptomatic cerebral hemorrhage was higher in the fibrinolytic treatment group than it was in the control group, however, 12% versus 4%, at 10 days. Mortality did not differ significantly: 24% in the treated group versus 27% in the control group. The patients all had intra-arterial treatment begun within 6 hours, with a mean time of symptom onset to treatment of 5.3 hours.
This important trial was successful and extended the therapeutic window to 6 hours. However, pro-UK is not commercially available, and it is not likely it will be. Substituting t-PA for pro-UK is not straightforward, because these agents may differ in optimal dosage and risk for intracerebral hemorrhage. 15 Clinical studies of safety, feasibility, and efficacy for intra-arterial interventions are imperative.

FUTURE DEVELOPMENTS: PHYSIOLOGICAL IMAGING FOR PATIENT TRIAGE
The imaging-based triage of patients with acute stroke is not new: all IV or intra-arterial intervention protocols require at least a noncontrast CT scan to exclude patients with intracerebral hemorrhage. There are three basic issues that are under investigation 16 : hemorrhage risk with fibrinolytic therapy, tissue viability, and vascular anatomy. The potential advantage for all of these approaches is to target a population most likely to benefit from a given intervention based on the underlying physiology. A specific intervention aimed at a physiologically homogenous group of patients is much more likely to be proved effective in a clinical trial.
Severely ischemic tissue has a greater risk of hemorrhage with lytic therapy than tissue with higher levels of cerebral blood flow. 17, 18 The risk with fibrinolytic treatment may be greater than the benefit for these patients, and exclusion of these patients from future clinical trials of revascularization may allow a better chance for showing a benefit with therapy. Clinical methods for the quantitative measurement of cerebral blood flow include single photon emission computed tomography (SPECT), positron emission tomography (PET), and stable Xenon-CT (Xe-CT). CT and magnetic resonance (MR) methods are under development.
The second issue under investigation is the prediction of tissue viability. The aim of this research is to target those likely to benefit and to extend the time window of intervention beyond 3 or 6 hours. In addition to having a high risk of hemorrhage with thrombolytic therapy, regions with very low levels of cerebral blood flow may also have little likelihood of recovery with recanalization. 18,19 Diffusion/perfusion MR also offers the potential for identifying those patients with viable tissue. 20 Most MR-based research involving diffusion and perfusion maps has been focused on this goal.
The final issue is vascular anatomy. Patients with acute thrombosis of the internal carotid artery may be less likely to benefit from IV therapy than will those with proximal middle or anterior cerebral artery emboli. 21 Similarly, those with proximal large artery occlusion may not do as well as those with more distal emboli. 22,23 For example, only 1 of 18 patients with a dense middle cerebral artery on initial CT has a good outcome with IV t-PA. 22 The larger the amount of thrombus, the poorer the result with IV t-PA. In the Acute Stroke Study Group, a pilot study for the NINDS t-PA trial, recanalization was found in only 9% (2 of 23) of patients with internal carotid artery occlusion, 26% (12 of 46) with middle cerebral artery (MCA) stem occlusion, and 44% (14 of 32) with MCA branch occlusions. 21,24 It is interesting to note that the recanalization rate for either MCA stem (M1) or branch (M2) occlusion with direct intra-arterial fibrinolysis was 62% after a 1-hour infusion of pro-UK. 8 In addition, patients with no visible large artery embolus may have nothing to gain from fibrinolysis and may benefit from pharmacological intervention to mitigate secondary injury or apoptosis. Twenty to 25% of patients presenting with acute ischemic stroke may have no visible arterial obstruction on emergent angiography performed within hours of presentation. 8,21,25

FUTURE DEVELOPMENTS: ENDOVASCULAR THERAPY
The potential for endovascular therapy in acute ischemic stroke is great. Direct delivery of lytic agents appears be more effective than IV infusions are in restoring flow. The rate of recanalization for M1 and M2 occlusion with pro-UK was 67% in the Pro-ACT II trial, 8 compared with 26 and 44%, respectively, for M1 and M2 occlusions treated with IV t-PA. 24 Mechanical devices for clot removal or dissolution offer the exciting possibility of more rapid restoration of flow, without the need for fibrinolytic agents and their added hemorrhagic risk. The development of image-based triage protocols may allow targeted therapy for patients most likely to benefit. A tremendous amount of work remains to be done, and many questions remain to be answered. With the exception of the Pro-ACT II study, all research in this field is at the level of feasibility and safety trials. In this section, we will review the data for the commercially available fibrinolytic agents and their recommended dosages, describe mechanical techniques of clot removal, and review different approaches for specific clinical scenarios, such as the postoperative patient and basilar artery or internal carotid artery thrombosis.

Intra-Arterial Fibrinolytic and Anti-Platelet Agents and Protocols
As discussed earlier, there are no fibrinolytic agents approved for intra-arterial use in the cerebrovasculature. There are no comparative data regarding different agents, doses, and rates of delivery. Table 1 summarizes nine different protocols. These protocols were developed by the Interventional Stroke Therapy Outcomes Registry (INSTOR) (principal investigator, R.R. [Buddy] Con-nors, M.D., http://www.strokeregistry.com). Figure 1 illustrates the case of a patient treated with the low-dose reteplase (Retevase™) protocol.
Alteplase (t-PA, Activase®) is a recombinant tissue plasminogen agent and has proved to be effective for the treatment of acute ischemic stroke as an IV infusion. However, it is insoluble in saline. In very dilute solutions, such as those proposed for intra-arterial infusion, it may precipitate and become less effective. Reteplase is a closely related recombinant t-PA with efficacy similar to that of alteplase for the treatment of acute coronary syndromes. It is water soluble, can be delivered as a bolus, and can be easily diluted. Many of the acute coronary intervention trials have used the combination of lowdose reteplase and an antiplatelet agent. 26 Antiplatelet agents also hold promise for the treatment of acute ischemic stroke. These can be used in a variety of ways: oral agents, IV infusions or bolus injections, and arterial injection and in combination with other fibrinolytic agents. Physiologically, antiplatelet agents can inhibit both platelet aggregation and platelet activation, both of which lead to the formation of additional thrombus. Abciximab (ReoPro®) inhibits platelet aggregation and activation by binding to the glycoprotein (GP) IIb/IIIa receptor site. It has been used in several coronary intervention trials to prevent rethrombosis after angioplasty and stenting and is associated with lower rates of intracranial hemorrhage than fibrinolytic agents are. 27 The rationale for the combination of a lowdose intra-arterial fibrinolytic agent and an IV bolus of a GP IIb/IIIa inhibitor, such as abciximab, is that revascularization will be accomplished with a reduction in the risk of intracerebral hemorrhage. Procedural thromboembolic events have also been successfully treated with IV or intra-arterial boluses of abciximab.

Mechanical Methods
Mechanical methods offer two major potential advantages over fibrinolytic therapy: speed of revascularization and reduced risk of secondary hemorrhage. Mechanical techniques under current investigation include snares, angioplasty (with or without a stent), and other devices to disrupt or dissolve the thrombus. Experience with these methods is anecdotal or in the form of small case series. 28,29 The data are promising, however.

Specific Scenarios
Basilar thrombosis is a difficult condition to diagnose and poses therapeutic problems as well. The clinical presentation for these patients may involve hours or days of stuttering events before a catastrophic loss of function. The outcome for these patients appears to be very poor, with mortality rates reported up to 100%. The length of time after the first ischemic event generally disqualifies most of these patients from IV fibrinolysis. Several large case series have suggested better outcomes than natural history with intra-arterial fibrinolysis. 30 The treatment of patients with complete occlusion of the internal carotid artery is also difficult because of problems with access, underlying disease, and clot burden. This is a common problem: a large proportion of the patients undergoing angiography in the Pro-ACT II trial had complete internal carotid artery occlusion. 8 The endovascular treatment of these patients depends primarily on anatomy. If there is good collateral flow and no apparent intracranial embolus, it may be best to leave things alone. If there is a distal embolus and good flow across the anterior or posterior commu-   nicating artery, direct lysis of the embolus via the patent collateral or the thrombosed internal carotid artery may be better than trying to recanalize the entire carotid. 31 Caution should be exercised when an underlying stenosis is found after fibrinolysis. The risk of reperfusionrelated hemorrhage with angioplasty in the setting of acute ischemia and recent fibrinolytic therapy may be very high. 15 This may be a particular problem if the flow to the distal tissue was chronically reduced prior to the stroke.
Patients with recent surgery are naturally excluded from treatment with an IV fibrinolytic agent. Intraarterial therapy can be accomplished with much lower doses and with relatively low risk. 32 These patients represent a population in which intra-arterial intervention within the first 3 hours after stroke can be easily justified.

CURRENT STATUS AND CONCLUSIONS
At present, IV thrombolysis with t-PA is the standard of care for patients presenting within 3 hours of onset. Patients presenting within 3 hours but with contraindications for IV therapy and those presenting between 3 to 6 hours can be offered intra-arterial fibrinolytic therapy with some scientific justification. However, it must be recognized that there are no data from randomized clinical trials regarding the efficacy of intra-arterial infusion with any commercially available fibrinolytic agent. This mode of treatment, therefore, remains investigational. The same is true for any of the mechanical methods for clot dissolution, for the application of imagingbased triage for patient selection or exclusion, and for the treatment of patients after 6 hours of onset. It is The anteroposterior right carotid injection shows M1 occlusion with pial collateralization to right MCA territory. (C) A 0.14 microcatheter to M1 shows occlusion. (D) After 0.05 units of retevase, partial flow was restored. (E) There was complete revascularization after 0.14 units. The patient made a complete recovery and was discharged the following day without neurological deficit. (F) Followup CT that morning showed blood or contrast in the right basal ganglia.
clear that acute stroke intervention is on the horizon. It is critical to develop Institutional Review Board-approved protocols for these investigational applications in order to provide adequate patient consent and to advance the field.