Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Intravenous (IV) recombinant tissue plasminogen activator (t-PA), commercially known as alteplase, remains the most commonly used therapy that is effective and safe for acute ischemic stroke.
Pivotal trials in the 1990s demonstrated the beneficial effect of t-PA; subsequent trials have confirmed the initial results and extended the evidence of benefit to include key subgroups such as the very old and patients with very mild strokes, as well as to extend the treatment window to 4.5 hours.
IV thrombolytic therapy is cost-effective, actually providing a net cost saving to the health care system.
The benefits of IV thrombolysis are continuously being improved upon. Newer thrombolytic agents may prove to be even safer and more effective. Combination treatment with intra-arterial intervention is of proven benefit and now standard of care for select patients.
The original protocol for using IV alteplase must be followed scrupulously to obtain the safety and benefits seen in the trials described. The selection criteria, however, have evolved over time as more evidence has become available.
The adage “Time is brain” remains: all studies have concluded that earlier treatment results in better outcomes. However, advanced imaging techniques have emerged as a surrogate to traditional time windows, and thus treatment beyond 4.5 appears possible.
Thrombolysis offers the simplest and most direct treatment for thrombotic disorders, including ischemic strokes. Plasminogen activators produce clinical improvement in patients with coronary artery thrombosis, peripheral vascular disease, venous thrombosis, pulmonary embolism, and acute ischemic stroke. As first shown in the pivotal National Institutes of Neurological Disorders and Stroke (NINDS) study, intravenous (IV) tissue plasminogen activator (t-PA) improved the clinical outcome of all types of ischemic stroke (i.e., large-artery, embolic, and small-vessel or lacunar strokes) if treatment began within 3 hours of the onset of symptoms. Consequently, the US Food and Drug Administration (FDA) approved recombinant t-PA (rt-PA) for the treatment of acute ischemic strokes within 3 hours of onset, excluding patients with intracranial hemorrhage (ICH). Moreover, further research has proven the use of rt-PA in extended time windows beyond 3 hours to be efficacious and safe in selected patients.
To date, other IV agents have proved useful, but none is yet approved by the FDA for the treatment of ischemic stroke. In this chapter, we review the historical background of thrombolytic therapy in preclinical and clinical trials, summarize the different agents in previous as well as current use, and discuss the management protocol for thrombolytic treatment in patients with stroke.
Thrombosis involves the processes of endothelial injury, platelet adherence and aggregation, and thrombin generation. Thrombin plays a major role in clot formation; it is responsible for cleaving fibrinogen to fibrin, which forms the clot matrix. Thrombin also activates factor XIII, which accomplishes interfibrin cross-linking. Fig. 53.1 illustrates the coagulation cascade. In a process involving platelet membrane receptors and phospholipids, thrombin is generated locally by the extrinsic and intrinsic pathways. Factors V and XIII interact with specific platelet membrane phospholipids to facilitate activation of factor X to factor Xa and the conversion of prothrombin to thrombin on the platelet surface. Platelet-bound thrombin-modified factor V (factor Va) serves as a high-affinity platelet receptor for factor Xa, which accelerates the rate of thrombin generation. The relative platelet–fibrin composition of a specific thrombus depends on regional blood flow or shear stress. At arterial flow rates, thrombi are predominantly platelet-rich, whereas at lower venous flow rates, the activation of coagulation predominates. The efficacy of thrombolysis perhaps depends on the relative fibrin content and fibrin cross-linking, the latter possibly determined by the age of the thrombus.
In addition to both endothelial-cell-derived antithrombotic characteristics and circulating anticoagulants (activated protein C and protein S), thrombus growth is limited by the endogenous thrombolytic system, in which plasmin plays a central role. One effect of endogenous thrombolysis is continuous remodeling of the thrombus. This effect results from the preferential conversion of plasminogen to plasmin on the thrombus surface, where fibrin binds t-PA in proximity to its substrate plasminogen, accelerating plasmin formation. Plasminogen activation may also occur on cells that express plasminogen receptors and produce plasminogen activators, such as endothelial and polymorphonuclear cells. If sufficient quantities of plasminogen activators are produced or administered, plasminogen can be activated in plasma, where it cleaves circulating fibrinogen and fibrin to produce fibrin split products.
The naturally circulating plasminogen activators, t-PA and single-chain urokinase-type PA (scu-PA), catalyze plasmin formation from plasminogen. In the circulation, plasmin rapidly binds to its inhibitor, α 2 -antiplasmin, and is inactivated. Endogenous fibrinolysis is modulated by several inhibitors of plasmin and plasminogen activators. The half-life of plasmin in the circulation is estimated to be approximately 0.1 second. α 2 -Antiplasmin is the primary inhibitor of fibrinolysis through plasmin inhibition by binding to excessive plasmin. Thrombospondin interferes with the t-PA-mediated, fibrin-associated activation of plasminogen. Contact activation inhibitors and C1 inhibitor have indirect effects on thrombolysis.
A competitive inhibitor of plasminogen is histidine-rich glycoprotein (HRG). In addition to inhibitors of plasmin, there are specific plasminogen activator inhibitors that decrease the activity of t-PA, scu-PA, and urokinase (UK) plasminogen activator (u-PA). Both plasma t-PA and u-PA are inhibited by plasminogen activator inhibitor-1 (PAI-1), which is derived from platelets and endothelial cells. The potential risk for thrombosis reflects the relative concentrations of circulating PAI-1 and the endogenous plasminogen activators t-PA and u-PA. In addition, other plasminogen activator inhibitors are derived from different tissues. Within the thrombus, however, plasmin is protected from this inhibitor, and t-PA is also relatively protected from circulating plasma inhibitors. This is why plasmin and t-PA can achieve their fibrinolytic effect better within the clot than in serum and also why clot lysis can be achieved with a relatively low risk of bleeding when these agents are used. Plasminogen activation is enhanced further by the complex formed by t-PA, fibrin, and plasminogen. The complex increases the clot-selective fibrinolytic activity of t-PA. Fibrinolysis occurs predominantly within the thrombus and at its surface. Lysis of thrombus is augmented by contributions from local blood flow. During thrombus consolidation, plasminogen bound to fibrin and to platelets allows local release of plasmin. Within the circulation, plasmin cleaves the fibrinogen to different fragments, which incorporate into the fibrin and cause destabilization of its network, therefore allowing further degradation.
All thrombolytic agents in current use are obligate plasminogen activators that act on fibrin and thrombin. Current thrombolytic agents are either endogenous plasminogen activators, which are involved in physiologic fibrinolysis, or exogenous plasminogen activators, which are not.
Tissue-type plasminogen activator is a single-chain, 70-kilodalton (kDa), glycosylated serine protease. It has four domains: finger or F domain, growth factor or E domain, two kringle regions (K1 and K2), and a serine protease domain. The COOH-terminal serine protease domain has the active site for the cleavage of plasminogen. The two kringle domains of t-PA ( Fig. 53.2 ) are similar to the kringle domains on plasminogen. The finger domain residues and the K2 domain residues are responsible for fibrin affinity. The single t-PA chain is converted to the double-chain t-PA form by plasmin cleavage of the arginine (position 275)–isoleucine (position 276) bond. Both the single- and double-chain forms are enzymatically active and have fibrin-selective properties. The plasma half-life of the single- and double-chain forms is 3–8 minutes. Tissue-type plasminogen is secreted by endothelial cells, neurons, astrocytes, and microglia. It is cleared by the liver. It is considered to be fibrin-dependent because of favorable activation of plasminogen in association with fibrin. Exercise and certain vasoactive substances, such as desmopressin, raise t-PA levels. Heparin and heparan sulfate increase t-PA activity. Recombinant DNA techniques are used to produce rt-PA for commercial use in both single-chain (alteplase) and double-chain (duteplase) forms. Fig. 53.2 illustrates the amino acid sequence of t-PA.
u-PA and its precursor scu-PA, or pro-UK, are glycoproteins. UK is synthesized by endothelial, renal, and malignant cells. The single-chain pro-UK possesses fibrin-selective plasmin-generating activity. Pro-UK has been synthesized by recombinant techniques to be used as an exogenous agent. Removal of the amino acid lysine at position 158 from scu-PA by plasmin produces the high-molecular-weight (HMW) double-chain u-PA (54 kDa) linked by the disulfide bridge; further cleavage produces the low-molecular-weight (LMW) u-PA (31 kDa). Both LMW and HMW forms are enzymatically active. HMW u-PA activates plasminogen to plasmin directly. The half-life of the two forms is 9–12 minutes. Pro-UK has been studied in patients with stroke but has not been approved for clinical use.
Different mutant forms of t-PA and u-PA have been developed through alteration of the original amino acid sequences by point mutations and deletions. These changes alter the specificity and stability of the molecules. A good example is tenecteplase—a mutant form of t-PA with delayed clearance and a longer half-life than t-PA. In patients with myocardial infarction (MI), tenecteplase has a half-life of 17 ± 7 minutes, as compared with 3.5 ± 1.4 minutes for alteplase. Tenecteplase has higher fibrin selectivity and greater resistance to plasminogen-activator inhibitor with enhanced lytic activity on the thrombus and induces earlier reperfusion than t-PA. The tenecteplase molecule is produced through alteration of the amino acid sequence at the T, N, and K domains of t-PA, as portrayed in Fig. 53.3 , resulting in the improved characteristics already described. Tenecteplase, as will be described later in this chapter, is currently garnering significant promise as it has been proven to be effective with a single bolus, simplifying acute thrombolysis protocols, and is currently under investigation in pivotal phase 3 clinical trials.
A recombinant form of human microplasminogen has also been investigated. Whereas t-PA is a specific proteolytic enzyme that converts the inactive proenzyme plasminogen to plasmin, microplasmin is a truncated form of plasmin, which has been tested in rodent models of ischemic stroke for safety and neuroprotective properties. It has been shown that in mice with inactivation of genes encoding α 2 -antiplasmin, this inactivation significantly reduced infarct size after ischemia, which suggests that there may be some neuroprotective properties inherent in the molecule. Microplasmin reacts with α 2 -antiplasmin and neutralizes it. Microplasmin was tested in two rabbit clot embolic stroke models, both small and large, with escalating weight-based dosing. Microplasmin improved behavioral rating scores 60 minutes after embolization without increasing hemorrhagic conversion. There have been no human studies to date.
Exogenous plasminogen activators are produced or extracted from nonhuman sources. Pharmacologic quantities of endogenous plasminogen activators produced by recombinant techniques, such as rt-PA, or produced through different mutations in the original physiologic plasminogen activator molecules, such as tenecteplase, have already been discussed.
Streptokinase (SK) is a single-chain polypeptide derived from group C β-hemolytic streptococci. SK combines with plasminogen, and the complex activates circulating plasminogen to plasmin and undergoes conversion to SK–plasmin itself. This complex is not inhibited by α 2 -antiplasmin, but SK activity can be eliminated by the presence of SK-neutralizing antibodies produced after previous infection with streptococci. The kinetics of SK elimination are complex, consisting of an initial half-life of 4 minutes and a second half-life of 30 minutes.
The recombinant plasminogen activators that are identical to the ones derived from the saliva of the vampire bat (Desmodus rotundus) include an alpha form that is more fibrin-dependent than t-PA. Its half-life is also longer than that of t-PA. Experimental studies have shown that the recombinant α-1 form and the bat plasminogen activator may be superior to t-PA in sustaining recanalization and may cause less fibrinogenolysis. The Desmoteplase in Acute Ischemic Stroke Study (DIAS) was a dose-finding randomized, phase 2 trial designed to evaluate the safety and efficacy of IV desmoteplase. Fixed doses of desmoteplase were evaluated, but the evaluation was terminated early due to the excessive rate of symptomatic ICH, defined as a four-point or more worsening of the NIHSS with computed topography-confirmed ICH. Subsequently, the lower weight-adjusted doses were investigated in 57 patients. A significantly higher rate of reperfusion was observed with desmoteplase compared with placebo ( P = .0012).
The Dose Escalation of Desmoteplase in Acute Stroke (DEDAS) study accompanied the DIAS study to further define the safety and efficacy of IV desmoteplase in patients with perfusion/diffusion mismatch 3–9 hours after stroke onset. DEDAS was a randomized, placebo-controlled, body-weight-adjusted dose-escalation study. Included were patients with an NIHSS score of 4 through 20, with a perfusion/diffusion mismatch on magnetic resonance imaging (MRI). The primary safety end point was symptomatic ICH defined as in the DIAS study. The study randomly assigned 37 patients: 14 received an IV bolus of 90 μg/kg of desmoteplase over 1–2 minutes, and 15 received 125 μg/kg. As was seen in the DIAS trial, reperfusion was found to correlate with a good clinical outcome at 90 days compared with those without reperfusion ( P = 0.003). Also, no symptomatic ICHs were observed.
Subsequently, DIAS 2, DIAS 3, DIAS-J (Japan), and DIAS 4 trials hoped to crystallize the safety and efficacy of desmoteplase in an extended time window. Unfortunately, no trials were able to demonstrate a significant clinical benefit of desmoteplase administered 3–9 hours after stroke in patients with major artery occlusion. Post hoc pooled analysis of all concomitant desmoteplase trials did indicate that a treatment dose of 90 μg/kg is safe and improves arterial recanalization; however, no functional outcome benefit was observed at 3 months post treatment. , Thus desmoteplase has largely been abandoned as a potential clinical thrombolytic agent.
Ancrod is a purified fraction of venom from the Malaysian pit viper (Calloselasma rhodostoma) and induces rapid defibrinogenation in humans by splitting fibrinopeptide A from fibrinogen. This agent has been the target for acute ischemic stroke since the 1980s. , Ancrod is given as a continuous infusion for up to 72 hours, and fibrinogen levels are checked before treatment and at designated intervals during and after treatment to determine the activity. Ancrod’s effects on plasma fibrinogen levels can be measured. The dosing strategy is to maintain a target fibrinogen level throughout the 5–7 days of dosing.
Early studies were promising; however, subsequent larger studies were halted due to futility. The North American Stroke Treatment with Ancrod Trial (STAT) was a randomized clinical trial in which patients received IV ancrod or placebo continuously for 72 hours, then intermittently for 2 days so that a target fibrinogen level could be obtained, which was based on their pretreatment fibrinogen level. This study was found to increase the proportion of patients with favorable functional status and similar mortality rates were seen between ancrod treatment and placebo. However, significantly more asymptomatic ICHs were also found in the ancrod group than in the placebo group. Following this study was the European Stroke Treatment with Ancrod Trial (ESTAT), which administered treatment within 6 hours of acute stroke onset. ESTAT was a randomized, double-blind placebo-controlled, phase 3 trial of 1222 patients randomly assigned to ancrod or placebo. Functional outcome at 3 months was similar in the ancrod and placebo groups. Symptomatic ICH occurred significantly more often in patients given ancrod compared with those given placebo ( P = .007).
A pivotal phase 3 trial was organized to confirm whether ancrod significantly altered the outcome after stroke in a large cohort of patients. The dose was carefully titrated to fibrinogen. Unfortunately, the trial was halted after an interim analysis for futility, which suggests that ancrod does not benefit patients with acute stroke. A Cochrane review of fibrinogen-depleting agents (six trials involving ancrod and two trials including defibrase) published in 2012 had similar findings. A total of 5701 patients were included across the 8 trials, and while there were fewer stroke recurrences in the treatment group than in the control group (relative risk [RR], 0.67; 95% confidence interval [CI], 0.49–0.92; 2 P = .01), symptomatic ICH was about twice as common in the treatment group compared with the control group (RR, 2.42; 95% CI, 1.65–3.56; 2 P < .00001). There are no further studies planned.
Considerable preclinical development showed that thrombolysis might be an effective stroke therapy. After recombinant technology was developed to produce large quantities of t-PA, animal studies could be conducted to show that t-PA, administered immediately after experimental embolic occlusion, caused reperfusion with significantly less neurologic damage. This development helped overcome the negative experience of early human use that accumulated before modern imaging techniques.
As early as 1963, Meyer et al. studied embolic stroke models in cats and monkeys and administered IV or intra-arterial bovine or human plasmin; this treatment resulted in clot lysis without higher rates of hemorrhagic infarction. In 1986, del Zoppo et al. demonstrated in baboons that after 3 hours of reversible balloon inflation compressing the middle cerebral artery (MCA), intracarotid administration of UK improved neurologic function and reduced infarct size without an increase in the rate of ICH detectable by CT. In 1985, Zivin et al. documented that t-PA could substantially improve neurologic function after embolization with artificially made clots. These studies together strongly suggested that thrombolysis, by restoring blood flow soon after stroke onset, could prevent neurologic deficits.
Preclinical trials also yielded insights into the potential risks of thrombolysis. In 1986, del Zoppo et al. studied t-PA-induced hemorrhagic transformation of ischemic baboon brains within 3.5 hours of MCA occlusion followed by 30 minutes of reperfusion. There was no significant difference in incidence or volume of infarct-related hemorrhage between any of the t-PA groups and the control group. In 1987, Slivka and Pulsinelli investigated the hemorrhagic potential of both t-PA and SK given 24 hours after experimental strokes in rabbits, as well as that of SK given 1 hour after experimental stroke. These investigators found that the thrombolytic agents increased the risk of ICH unless they were given early after the insult. In 1989, Lyden et al. found no difference in the frequency of hemorrhagic transformation in the ischemic brains of rabbits whether t-PA was administered 10 minutes, 8 hours, or even 24 hours after cerebral embolism. In 1991, Clark et al. demonstrated that aspirin and t-PA act synergistically to cause intracranial bleeding in the rabbit embolism model.
To learn whether hemorrhagic risk was associated with thrombolytic agents in general or with a particular agent specifically, Lyden et al. compared t-PA, SK, and saline given after embolic stroke in rabbits. SK, but not t-PA, was associated with a significant increase in ICH rate and size. Table 53.1 demonstrates those results. It should be noted that there was no clear dose-response effect for hemorrhages, and the doses used were comparable to those used for cardiac disease in humans. Only the rabbits in which t-PA achieved thrombolysis had twice the frequency of ICH than those given placebo, which suggests that reperfusion might be the basis for the higher rate of hemorrhagic transformation.
Treatment | Dose | Time (min) | n | Hemorrhage | Thrombolysis | ||
---|---|---|---|---|---|---|---|
No. | % | No. | % | ||||
Saline | b | 48 | 12 | 25 | 17 | 35 | |
t-PA | 3 mg/kg | 90 | 16 | 5 | 31 | 9 | 56 |
t-PA | 5 mg/kg | 90 | 22 | 3 | 14 | 15 | 68 |
t-PA | 10 mg/kg | 90 | 11 | 4 | 36 | 10 | 91 b |
SK | 30,000 units/kg | 5 | 11 | 6 | 55 | 5 | 45 |
SK | 30,000 units/kg | 90 | 17 | 11 | 65 c | 14 | 82 b |
SK | 30,000 units/kg | 300 | 12 | 10 | 83 c | 10 | 83 b |
a Results of t-PA treatment at 5 minutes and 4, 8, and 24 hours are contained in references 2 and 3.
b Saline-treated control rabbits were treated 5, 90, or 300 minutes after embolization.
In summary, preclinical studies suggested that t-PA had reliably opened cerebral arteries in embolic experimental models. Considerable benefit was achieved if thrombolysis occurred early after occlusion onset. Hemorrhages occurred after thrombolysis and seemed to be related to the particular agent used, and SK carried a greater risk than t-PA.
The clinical development of thrombolysis for stroke proceeded logically from preclinical testing. Early experiments benefited from preclinical data and emphasized several factors: agent, dose, timing, and concomitant management. We review first human-use studies that documented thrombolysis in humans after the administration of thrombolytic agents.
Dose-ranging studies yielded important data about the dose of t-PA to use in pivotal trials; the efficacy of the agents seemed to be counterbalanced by hemorrhages at higher doses. Large placebo-controlled trials confirmed the efficacy and hazards of these agents, as well as observations from preclinical studies that SK was more hazardous. Finally, after regulatory approval of t-PA for treatment of acute stroke, open-label studies confirmed the findings of the definitive trials and showed that IV thrombolysis is feasible and efficacious in a variety of settings. Data from experimental cerebral ischemia studies pointed to the need to treat acute stroke as soon as possible, and this observation also proved true in human trials.
Results of early attempts to achieve thrombolysis for acute ischemic stroke were discouraging, especially in studies conducted without the benefit of CT imaging to exclude hemorrhage. Additionally, in these preliminary trials, patients were enrolled within significantly longer time windows than currently approved. In 1965, Meyer et al. studied 73 patients with acute progressive strokes; the treatment group received SK plus anticoagulation, and the control group received anticoagulation only. There was a higher incidence of death in the treatment group and better clinical improvement in the control group.
In 1976, Fletcher et al. studied 31 patients with acute ischemic stroke who were treated with one of three different doses of IV UK; treatment was given within 36 hours of symptom onset. The study concluded that UK could be administered to patients in doses that achieve substantial thrombolysis without producing other than mild coagulation deficits; this study could not address the efficacy of the treatment, however, because the number of patients was too low. The mortality rate was 16%, and there was no placebo group for comparison. On the basis of these two studies, which were widely discussed, IV thrombolysis for stroke was abandoned pending better agents and better selection procedures.
After the efficacy and safety of t-PA were proved in animal models, thrombolysis was pursued again in acute clinical stroke trials. In 1992, del Zoppo et al. studied 139 patients with acute ischemic stroke who received different doses of IV t-PA within 8 hours of stroke onset. An angiogram confirmed occlusion of an extracranial or intracranial arterial cerebral blood supply in all patients. Exclusion criteria included a minor deficit, a transient ischemic attack (TIA), a clinically large stroke with a combination of hemiplegia, impaired consciousness, and forced gaze deviation, blood pressure higher than 200 mm Hg systolic, 120 mm Hg diastolic, and radiologic (CT) evidence of bleeding or radiologic evidence of significant mass effect or midline shift. Patients with early CT hypoattenuation changes were not excluded from the study. Primary end points were angiographic recanalization and ICH with neurologic deterioration. This landmark study re-established the clinical promise of thrombolysis; 40% of all patients experienced recanalization of occluded arteries. Intriguingly, there was no relation between dose and recanalization, but patients with distal (i.e., smaller) clots showed higher recanalization rates. The frequency of all hemorrhages was 30.8%, although symptomatic hemorrhages occurred in 9.6% of all patients. The mortality rate during hospitalization was 12.5%. There was no increase in hemorrhages with doses comparable to those used to achieve coronary reperfusion, although it could not be assumed that the safe and effective dose for acute coronary events would be the perfect dose for acute stroke treatment. Therefore, the effective and safe dose for stroke treatment was yet to be determined.
In 1992, the first in a series of government-sponsored trials appeared. In a dose-finding trial sponsored by the NINDS, 74 patients with acute ischemic stroke received escalating doses of t-PA (0.35–1.08 mg/kg) within the time window of 90 minutes. Intracranial hematomas did not occur in any of the 58 patients who received doses of 0.85 mg/kg or less, but did occur with higher doses. Hemorrhages associated with neurologic deterioration (symptomatic hemorrhages) occurred in 3 of the 74 patients, although such hemorrhages did not occur at t-PA doses of less than 0.95 mg/kg. Major improvement, manifesting as a significant improvement in the NIHSS score, occurred at 2 hours in 30% of the patients and at 24 hours in 46% of patients. Major neurologic improvement was not related to the dose of t-PA. The investigators concluded that the highest safe dose of t-PA was probably less than 0.95 mg/kg, leading to standard-of-care dosing of 0.9 mg/kg that is still used today. Many years later, Diedler et al. investigated the safety and efficacy of the accepted maximum dose of 90 mg in patients weighing greater than 100 kg, who by definition receive a lower per-kilogram dose compared with patients weighing less than 100 kg. Interestingly, there was a higher incidence of symptomatic ICH in patients greater than 100 kg, while major neurologic improvement and functional independence were similar. Their conclusion supports the current upper dose limit.
In 1992, Haley et al. studied 20 patients with acute ischemic stroke in another dose-escalating trial in which t-PA treatment was given between 91 and 180 minutes after stroke onset. The risks of symptomatic ICH were 10% overall and 17% with the two higher dosage levels (the three doses used were 0.6 mg/kg, 0.85 mg/kg, and 0.95 mg/kg). Three patients (15%) improved by four points on the NIHSS at 24 hours.
Mori et al. conducted a trial in Japan in which either 6 million or 12 million units of IV t-PA or placebo were administered within 6 hours of stroke onset. Using angiograms before and after thrombolysis, these investigators confirmed that t-PA increased the rate of MCA recanalization. Of considerable importance is the fact that functional outcome measured by the BI score was also significantly improved by thrombolysis. Like the del Zoppo trial, this trial established unequivocally that IV thrombolytics could open occluded cerebral vessels. Further, and perhaps even more important, the trial results suggested that angiographic confirmation of cerebral vessel occlusion might not be essential before IV thrombolysis.
In 1993, in the “bridging trial,” a forerunner of the definitive NINDS study, Haley et al. studied 27 patients who received 0.85 mg/kg of IV t-PA or placebo within 3 hours of stroke onset. This was a randomized, double-blinded, placebo-controlled study. Despite the small sample size, there was a suggestion of early neurologic improvement (at 24 hours) in the patients treated with t-PA. In the treatment arm in which therapy was given up to 90 minutes after stroke onset, 6 of the 10 patients who received t-PA improved by four or more points on the NIHSS, compared with 1 of the 10 patients given placebo. In the treatment arm in which therapy was given between 91 and 180 minutes after stroke onset, two patients in the t-PA group and two patients from the placebo subgroups improved by four or more points on the NIHSS at 24 hours. The results of the bridging trial anticipated those of the larger NINDS study in a surprising number of respects. Nevertheless, large, rigorous, placebo-controlled, randomized trials were needed to confirm any beneficial effects afforded by IV thrombolytic agents.
Published in 1995, the European Cooperative Acute Stroke Study (ECASS) included 620 patients treated with 1.1 mg/kg of IV t-PA or placebo within 6 hours of stroke onset. The trial showed no significant efficacy in the intent-to-treat primary analysis. On exclusion of patients with protocol violations (109 patients, 17.4%), a target population of 511 patients was selected for further analysis. Protocol violations consisted of inclusion of patients with large strokes (i.e., hypodensity of greater than one-third of the MCA territory on CT), concurrent use of anticoagulants or volume expanders, detection of hemorrhage on baseline CT, uncontrolled hypertension, and lack of complete follow-up. The first hypothesis in this study was that there would be a 15-point difference in the BI between the two groups in the study, favoring the t-PA treatment group. The second hypothesis was that there would be a difference on the mRS in favor of the t-PA group.
In the target population, there was a one-point difference in the mRS score between the two groups ( P = .035) in favor of the t-PA group. There was no statistically significant difference in ICH rates between the groups, but there was an increase in frequency of large parenchymal hemorrhages in the t-PA group and an increase in frequency of hemorrhagic infarcts in the placebo group. There was no statistically significant difference in mortality rates at 30 days. Although ECASS failed to show a benefit (the hypothesis was not proved), subsequent analyses showed a significant treatment effect. In particular, on post hoc reanalysis of ECASS with the use of NINDS global end point statistics, a statistically significant treatment effect was detected in the intent-to-treat group. This finding suggests that ECASS might have shown a beneficial effect of thrombolytic agents in stroke, even though one cannot definitely reach that conclusion from a post hoc analysis. Furthermore, when the patients treated within 3 hours were examined separately (38 given placebo, 49 given t-PA), a non-statistically significant treatment effect was demonstrated by the same statistical analysis methods used in the NINDS study (global odds ratio [OR], 2.3; P = .07). The ECASS post hoc analyses suggested that an independent 3-hour trial might show a benefit for thrombolytic agents.
In December 1995, the NINDS study was published; it was a randomized, placebo-controlled, multicenter trial that showed the efficacy of t-PA in treating acute ischemic strokes within 3 hours of onset. This NINDS study differed from ECASS in several respects besides the dose of t-PA and time to treatment. Most importantly, NINDS protocol required that the blood pressure had to be controlled to below 185 mm Hg systolic, 95 mm Hg diastolic. Box 53.1 summarizes the inclusion and exclusion criteria of the original NINDS study, which has changed significantly over the years and will be discussed in detail later in the chapter.
Ischemic stroke of defined onset <3 h
Deficit measurable on NIHSS
Baseline CT of the brain without evidence of hemorrhage
A prior stroke within the last 3 months PTP
Major surgery within the last 14 days PTP
Serious head trauma within the last 3 months PTP
History of ICH
Systolic BP >185 mm Hg or diastolic BP >110 mm Hg or if aggressive treatment was required to lower the BP to below these limits
Rapidly improving or minor symptoms
Symptoms suggestive of SAH
Gastrointestinal bleeding or urinary tract hemorrhage within 3 weeks PTP
Arterial puncture at a noncompressible site within the last 7 days PTP
Seizure at the onset of symptoms
Anticoagulants or heparin within 48 h before stroke onset or elevated PTT or elevated PT >15 s
Platelet count <100,000/mL
Blood glucose <50 mg/dL or above 400 mg/dL
The NINDS study had two parts with identical protocols but different end points. Part 1 tested whether t-PA showed clinical activity, as indicated by a statistically significant difference on the primary end point, chosen arbitrarily to be either an improvement of 4 or more points on the NIHSS or complete resolution of the neurologic deficit within 24 hours. Part 2 used a global test statistic to assess clinical outcome after 3 months, based on scores on the BI, mRS, Glasgow Outcome Scale (GOS), and NIHSS. Part 1 enrolled 291 patients (144 in the t-PA group and 147 in the placebo group), and part 2 enrolled 333 patients (168 patients in the t-PA group and 165 patients in the placebo group). In part 1, on the primary end point, the number of patients improving by four or more points on the NIHSS at 24 hours was 67 (47%) in the t-PA group and 57 (39%) in the placebo group (not statistically significant, with a P value of .21). Subsequent analysis showed that any other cutoff improvement in the 24-hour NIHSS score, such as five or more points, would have yielded a statistically significant difference between the two groups ( Fig. 53.4 ).
In part 2 of the NINDS, benefit was observed on all four primary efficacy measures (i.e., NIHSS, BI, mRS, and GOS scores) at 3 months from onset of stroke. Patients treated with t-PA were 30%–50% more likely to have minimal or no disability at 3 months, depending on the outcome measure. For example, the percentage of patients with an mRS score of 1 or less at 3 months was 39% in the t-PA group versus 26% in the placebo group (OR, 1.7; 95% CI, 1.1–2.6; P = .019). Symptomatic ICH occurred in 6.4% of patients who received treatment but only in 0.6% of patients who received placebo. Mortality rates at 3 months were not statistically different between the two groups, being 17% in the t-PA group and 21% in the placebo group. Thus, despite an increased risk of hemorrhage, the mortality rate was not affected, and IV t-PA provided considerable benefit and improved outcomes, as depicted in Fig. 53.5 . Furthermore, the NINDS data analysis showed that t-PA treatment resulted in a more favorable outcome regardless of the subtype of stroke (small-vessel, large-vessel, or cardioembolic stroke) diagnosed at baseline.
Further subgroup analysis of the NINDS data showed that the only variables independently associated with an increased risk of symptomatic ICH in the t-PA-treated patients were the baseline severity of the stroke as measured by the NIHSS, brain edema defined by hypodensity on baseline CT, and mass effect on baseline CT (before treatment). These factors did not interact with treatment, however, which suggests that such factors might not be predictive in excluding patients from treatment.
Subsequent prespecified analyses of the NINDS database with the use of a global statistical method showed sustained, statistically significant benefit at 6-month and 1-year follow-up points: the OR values for a favorable outcome in the t-PA group compared with the placebo group were 1.7 with a 95% CI of 1.3–2.3 at 6 months and 1.7 with a 95% CI of 1.2–2.3 at 1 year. At 1 year, the range of absolute increase in the percentage of patients with a favorable outcome was 11%–13%, and the range of relative increase in the percentage of patients with a favorable outcome was 32%–46% for the three outcome scales (mRS, BI, and GOS). Patients treated with t-PA were at least 30% more likely to be independent at 1 year than those given placebo. Importantly, favorable outcomes were not accompanied by an increase in severe disability or mortality. The proportion of patients surviving between 3 months and 12 months after stroke was consistently higher in the t-PA group than in the placebo group. However, there was no statistically significant difference in mortality at 6 months and 1 year. After adjustment for those variables, treatment with t-PA still offered better outcomes.
As was the case for the 3-month follow-up data, there was no evidence of interaction between the subtype of stroke at baseline and treatment, meaning that all stroke subtypes (large-vessel, small-vessel, and embolic) benefitted from t-PA. Moreover, there was no significant difference in the incidence of recurrent stroke between the t-PA and placebo groups at 1 year. Furthermore, another analysis of the NINDS data addressed finding the binary measures that predicted effectiveness of t-PA during the first 3 months. Measures using NIHSS and mRS scores of 1 or less were the most sensitive discriminators of effectiveness of t-PA in the NINDS study. The best measure was NIHSS score of 2 or less at 24 hours. High-quality analysis of the volume of brain infarction as measured by CT was not as sensitive in detecting a treatment effect as the clinical scale measures.
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here