- Email: [email protected]
For how long is brain tissue salvageable? Thrombolysis-based evidence
For How Long Is Brain Tissue Salvageable? Thrombolysis-Based Evidence Lawrence Wechsler, md, and Patrick Lyden, md
Recombinant t-PA (tissue plasminogen activator, alteplase) was licensed for use in the United States for treatment of acute ischemic stroke on the basis of the National Institute of Neurological Disorders and Stroke (NINDS) study. In this trial, t-PA was started within 3 hours of onset of stroke symptoms. Two other trials of t-PA (the European Cooperative Acute Stroke Study [ECASS] I and ECASS II) failed to detect a significant effect in stroke patients. Both of these trials allowed treatment to begin up to 6 hours after the onset of symptoms. However, the results of a meta-analysis of all well-controlled trials of thrombolytics suggest that thrombolysis is an effective strategy when started up to 6 hours after the onset of symptoms. More than 50% of the trials included in this meta-analysis were, in fact, of t-PA. Allowing time for infusion of the thrombolytic and time for lysis of the thrombus, the findings of the meta-analysis suggest that salvageable tissue may remain for at least 8 hours after the onset of symptoms. Prourokinase was tested for efficacy in acute stroke in the PROlyse in Acute Cerebral Thromboembolism (PROACT) I trial. This trial had an insufficient number of patients to detect a meaningful effect. Consequently, the much larger PROACT II trial was conducted. In both of these studies, infusion began within 6 hours of the onset of symptoms. PROACT II demonstrated a significant improvement in outcome, and it can be concluded that, in some patients at least, salvageable tissue remained for at least 8 hours after the onset of symptoms. Consequently, treatment with a neuroprotectant during this 8-hour window might have been effective. Key Words: Stroke— Neuroprotection—Thrombolysis—t-PA—Excitotoxicity.
The basic tenet that underlies the development of neuroprotectant drugs for use in the treatment of ischemic stroke is that death is not instantaneous for all of the neurons in the territory supplied by the artery that becomes occluded. There is a central core of completely anoxic cells that will die rapidly. At the periphery of the ischemic volume, there is a shell of neurons that receive a sufficient collateral blood supply to ensure that they survive and possibly recover once reperfusion occurs.
From the Stroke Institute, University of Pittsburgh Medical Center, Pittsburgh, PA; and UCSD Medical Center and Veterans’ Affairs Medical Center, San Diego, CA. Address correspondence to Lawrence Wechsler, MD, Stroke Institute, University of Pittsburgh Medical Center, 200 Lothrop St, Pittsburgh, PA 15213. Copyright © 2000 by National Stroke Association 1052-3057/00/0906-0205$3.00/0 doi:10.1053/jscd.2000.19321
Between these 2 zones is an intermediate shell of neurons that are immutably destined neither to survive nor to die. These cells are the potential target of neuroprotectant drugs.1 As time passes, cells in this “penumbra” continue to die. The following question could therefore be posed: “For how long are at-risk neurons salvageable?” In reality, it is likely that some at-risk cells remain salvageable for many hours or even days. At some point, the remaining volume of tissue that could be protected from the excitotoxic cascade is too small to provide any clinically relevant functional benefit. The question could therefore be posed differently: “For how long after the onset of symptoms is it worth starting neuroprotectant therapy?” Some clues can be found by examining the information that has been obtained during trials of neuroprotectants, and these are reviewed in the article by Hsu et al. in this supplement (pp. 24-31). Baron and Moseley (pp. 15-20 in
Journal of Stroke and Cerebrovascular Diseases, Vol. 9, No. 6, Suppl 2 (November-December), 2000: pp 21-23
21
L. WECHSLER AND P. LYDEN
22
this supplement) review information about the time window from brain-imaging studies. This article focuses on information that can be gained from analysis of studies with thrombolytic agents.
Recombinant t-PA (Alteplase) The National Institute of Neurological Disorders and Stroke (NINDS) rt-PA (recombinant tissue plasminogen activator) Stroke Study Group conducted a trial in 624 patients with stroke. They found that 11% to 13% more patients in the t-PA group than in the placebo group had a good functional outcome.2 As a result, t-PA was approved in the United States for the treatment of acute ischemic stroke on the basis of treatment being started within 3 hours of the start of symptoms. The success of thrombolysis started within 3 hours of stroke onset suggests that a clinically relevant volume of at-risk but salvageable neurons survived until reperfusion started. The earliest that reperfusion can occur is at the beginning of the t-PA infusion, but it is more likely that reperfusion begins during the infusion (i.e., up to 4 hours after symptom onset) or shortly thereafter (perhaps up to 7 hours). Two similar trials have been conducted in Europe—the European Cooperative Acute Stroke Study (ECASS) I3 and ECASS II.4 Both ECASS I and ECASS II allowed treatment with t-PA to be started up to 6 hours after the onset of symptoms. ECASS I used a higher dose of t-PA than the NINDS study (1.1 mg/kg compared with 0.9 mg/kg) but failed to detect any significant advantage for the t-PA group over the placebo group, despite finding effect sizes comparable with those in the NINDS trial. In ECASS I, however, mortality and intracerebral hemorrhages were more common in the t-PA group than in the placebo group. ECASS II used the same dose of t-PA as the NINDS study (0.9 mg/kg). Like ECASS I, ECASS II failed to find any significant benefit of giving patients t-PA.4 Although the study stratified patients according to whether treatment was started less than 3 hours or between 3 and 6 hours after the onset of symptoms, most patients (approximately 80%) began to receive t-PA in the 3- to 6-hour window. A simplistic conclusion from the ECASS studies would be that there are no clinically relevant at-risk neurons that can be rescued if thrombolytic therapy is delayed until more than 3 hours after the onset of symptoms. Using the same assumption as before about the time to reperfusion, this translates to at least 4 to 8 hours after the onset of symptoms. Although the effect of starting thrombolytic treatment more than 6 hours after the onset of symptoms has not been tested, it is unlikely that there are viable at-risk cells after 8 hours if there are none left between 4 and 8 hours.
There are some reasons to conclude that the story is not so simple. The results of a meta-analysis suggest that thrombolytic therapy does have a beneficial effect when started up to 6 hours after the onset of symptoms.5 Approximately half of the data in this meta-analysis came from studies of t-PA. It is therefore possible that the ECASS studies contained a type 2 statistical error, finding no significant effect when in reality the drug did have a beneficial effect, even when started up to 6 hours after symptom onset. It has been pointed out, for example, by Bath,6 that patients in the ECASS studies, particularly in ECASS II, had milder strokes than those in the NINDS study. Patients with mild strokes are more likely to recover without treatment, and therefore the difference between treated and untreated patient groups will be smaller and may not reach statistical significance. The ECASS trials also used different criteria for success than NINDS. In the NINDS study, a global test statistic (the Wald test) was used to assess clinical outcome,7,8 combining scores on the Barthel index,9 the modified Rankin scale,10 the Glasgow outcome scale,11 and the National Institutes of Health Stroke Scale (NIHSS).12 ECASS II, however, used “favorable” outcome rather than “independence” as the primary outcome measure. A post hoc analysis of ECASS II that used independence as the outcome criterion found that t-PA was statistically superior to placebo—treatment with t-PA resulted in an 8.3% increase in the number of patients rated as independent at 90 days. If ECASS II contained a type 2 statistical error, then we may conclude that there may be viable but at-risk neurons as late as 8 hours after the onset of symptoms of ischemic stroke.
Prourokinase The recanalizing efficacy and safety of intra-arterial prourokinase was reported in the PROlyse in Acute Cerebral Thromboembolism (PROACT) I trial.13 This study was not powered to detect any effect of prourokinase on clinical outcome. Consequently, a larger, double-blind, placebo-controlled trial—the PROACT II trial—was performed in patients with acute ischemic stroke.14 In both studies, treatment with prourokinase was started within 6 hours of the onset of symptoms. As primary outcome criteria, the PROACT II investigators used the proportion of patients functioning independently at 90 days as defined by a modified Rankin score of 2 or less. The recruitment criteria were designed to be very demanding in an attempt to reduce the heterogeneity of the sample. Patients were included only with angiographically proven M1 or M2 MCA territory occlusions. This meant that 12,323 patients were screened to identify 180 who were eligible to be randomized to receive either prourokinase
SALVAGING BRAIN TISSUE: THROMBOLYSIS-BASED EVIDENCE
plus intravenous heparin (n ⫽ 121) or intravenous heparin alone (n ⫽ 59). The mean NIHSS score was 17, indicating that these patients had more severe strokes than those in the NINDS, ECASS I, or ECASS II studies. A further refinement of the trial was that patients were stratified according to their NIHSS scores on entry. Overall, the primary analysis indicated that 40% of prourokinase-treated patients were independent at 90 days (modified Rankin score ⱕ 2), compared with only 25% of the control group (P ⫽ .043). Angiographically proven partial or complete recanalization rates 2 hours after the start of the infusion were 66% for the prourokinase group and 18% for the placebo group, although the complete recanalization rates were only 19% and 2%, respectively. The design of the study did not allow the calculation of a mean time to recanalization but did report the median time to starting infusion as being 5.3 hours after stroke onset. Clearly, the recovery rate and recanalization rates (both partial and complete) differed—the rate of recovery was lower than the reperfusion rate. This implies that an appreciable fraction of the patients whose clots were successfully dissolved had no salvageable tissue at the time blood flow was reestablished. It is also clear that many patients did benefit and therefore must have had at-risk tissue that was saved. Taking into account a median time from symptom onset to starting treatment of approximately 5 hours and a 2-hour infusion period, it can be estimated that salvageable tissue remained for up to approximately 7 hours, in at least some patients. Because the figure of 5.3 hours is a median, a number of the patients will have benefited from treatment begun later than this. Nevertheless, using a conservative estimate of 5 hours, the following argument applies: if the ischemic tissue that was saved by thrombolysis was at risk of death from the excitotoxic cascade, the time at which it would have been worth starting neuroprotectant treatment may have extended to up to 7 hours after the onset of symptoms.
Conclusions The evidence from the NINDS study of t-PA indicates that ischemic stroke tissue may be salvageable for up to 5 hours after the onset of symptoms. The PROACT II study of prourokinase suggests a longer time window—a clinically relevant volume of salvageable tissue may remain for at least 7 hours after the onset of symptoms. If reperfusion began after the end of infusion, then there may indeed be viable and salvageable tissue remaining for more than 8 hours after the onset of symptoms in
23
some patients. If the quiescent but salvageable tissue in the ischemic territory is at risk of death because of the excitotoxic cascade, then treatment with thrombolytic plus neuroprotectant should result in the recovery of more tissue than treatment with thrombolytic alone. It is therefore reasonable to conclude from the evidence from studies with thrombolytics that the window for starting a neuroprotectant is at least 7 hours and possibly longer after the onset of stroke symptoms.
References 1. Olney JW. Excitotoxic amino acids and neuropsychiatric disorders. Annu Rev Pharmacol Toxicol 1990;30:47-71. 2. The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. Tissue plasminogen activator for acute ischemic stroke. N Engl J Med 1995; 333:1581-1587. 3. Hacke W, Kaste M, Fieschi C, et al. Intravenous thrombolysis with recombinant tissue plasminogen activator for acute hemispheric stroke. The European Cooperative Acute Stroke Study (ECASS). JAMA 1995;274:1017-1025. 4. Hacke W, Kaste M, Fieschi C, et al. Randomised doubleblind placebo-controlled trial of thrombolytic therapy with intravenous alteplase in acute ischaemic stroke (ECASS II). Second European–Australasian Acute Stroke Study Investigators. Lancet 1998;352:1245-1251. 5. Wardlaw JM, del Zoppo G, Yamaguchi T. Thrombolysis for acute ischaemic stroke. Cochrane Database Syst Rev 2000;2:CD000213. 6. Bath P. Alteplase not yet proven for acute ischaemic stroke [comment]. Lancet 1998;352:1238-1239. 7. Lefkopoulou M, Moore D, Ryan L. The analysis of multiple correlated binary outcomes: Application to rodent teratology experiments. J Am Stat Assoc 1989;84:810-815. 8. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986;42:121130. 9. Mahoney FI, Barthel DW. Functional evaluation: The Barthel Index. Md Med J 1965;14:61-65. 10. van Swieten JC, Koudstaal PJ, Visser MC, et al. Interobserver agreement for the assessment of handicap in stroke patients. Stroke 1988;19:604-607. 11. Teasdale G, Knill-Jones R, van der SJ. Observer variability in assessing impaired consciousness and coma. J Neurol Neurosurg Psychiatry 1978;41:603-610. 12. Lyden P, Brott T, Tilley B, et al. Improved reliability of the NIH Stroke Scale using video training. NINDS TPA Stroke Study Group. Stroke 1994;25:2220-2226. 13. del Zoppo GJ, Higashida RT, Furlan AJ, et al. PROACT: A phase II randomized trial of recombinant pro-urokinase by direct arterial delivery in acute middle cerebral artery stroke. PROACT Investigators. Prolyse in Acute Cerebral Thromboembolism. Stroke 1998;29:4-11. 14. Furlan A, Higashida R, Wechsler L, et al. Intra-arterial prourokinase for acute ischemic stroke. The PROACT II study: a randomized controlled trial. Prolyse in Acute Cerebral Thromboembolism. JAMA 1999;282:2003-2011.
Recombinant t-PA (tissue plasminogen activator, alteplase) was licensed for use in the United States for treatment of acute ischemic stroke on the basis of the National Institute of Neurological Disorders and Stroke (NINDS) study. In this trial, t-PA was started within 3 hours of onset of stroke symptoms. Two other trials of t-PA (the European Cooperative Acute Stroke Study [ECASS] I and ECASS II) failed to detect a significant effect in stroke patients. Both of these trials allowed treatment to begin up to 6 hours after the onset of symptoms. However, the results of a meta-analysis of all well-controlled trials of thrombolytics suggest that thrombolysis is an effective strategy when started up to 6 hours after the onset of symptoms. More than 50% of the trials included in this meta-analysis were, in fact, of t-PA. Allowing time for infusion of the thrombolytic and time for lysis of the thrombus, the findings of the meta-analysis suggest that salvageable tissue may remain for at least 8 hours after the onset of symptoms. Prourokinase was tested for efficacy in acute stroke in the PROlyse in Acute Cerebral Thromboembolism (PROACT) I trial. This trial had an insufficient number of patients to detect a meaningful effect. Consequently, the much larger PROACT II trial was conducted. In both of these studies, infusion began within 6 hours of the onset of symptoms. PROACT II demonstrated a significant improvement in outcome, and it can be concluded that, in some patients at least, salvageable tissue remained for at least 8 hours after the onset of symptoms. Consequently, treatment with a neuroprotectant during this 8-hour window might have been effective. Key Words: Stroke— Neuroprotection—Thrombolysis—t-PA—Excitotoxicity.
The basic tenet that underlies the development of neuroprotectant drugs for use in the treatment of ischemic stroke is that death is not instantaneous for all of the neurons in the territory supplied by the artery that becomes occluded. There is a central core of completely anoxic cells that will die rapidly. At the periphery of the ischemic volume, there is a shell of neurons that receive a sufficient collateral blood supply to ensure that they survive and possibly recover once reperfusion occurs.
From the Stroke Institute, University of Pittsburgh Medical Center, Pittsburgh, PA; and UCSD Medical Center and Veterans’ Affairs Medical Center, San Diego, CA. Address correspondence to Lawrence Wechsler, MD, Stroke Institute, University of Pittsburgh Medical Center, 200 Lothrop St, Pittsburgh, PA 15213. Copyright © 2000 by National Stroke Association 1052-3057/00/0906-0205$3.00/0 doi:10.1053/jscd.2000.19321
Between these 2 zones is an intermediate shell of neurons that are immutably destined neither to survive nor to die. These cells are the potential target of neuroprotectant drugs.1 As time passes, cells in this “penumbra” continue to die. The following question could therefore be posed: “For how long are at-risk neurons salvageable?” In reality, it is likely that some at-risk cells remain salvageable for many hours or even days. At some point, the remaining volume of tissue that could be protected from the excitotoxic cascade is too small to provide any clinically relevant functional benefit. The question could therefore be posed differently: “For how long after the onset of symptoms is it worth starting neuroprotectant therapy?” Some clues can be found by examining the information that has been obtained during trials of neuroprotectants, and these are reviewed in the article by Hsu et al. in this supplement (pp. 24-31). Baron and Moseley (pp. 15-20 in
Journal of Stroke and Cerebrovascular Diseases, Vol. 9, No. 6, Suppl 2 (November-December), 2000: pp 21-23
21
L. WECHSLER AND P. LYDEN
22
this supplement) review information about the time window from brain-imaging studies. This article focuses on information that can be gained from analysis of studies with thrombolytic agents.
Recombinant t-PA (Alteplase) The National Institute of Neurological Disorders and Stroke (NINDS) rt-PA (recombinant tissue plasminogen activator) Stroke Study Group conducted a trial in 624 patients with stroke. They found that 11% to 13% more patients in the t-PA group than in the placebo group had a good functional outcome.2 As a result, t-PA was approved in the United States for the treatment of acute ischemic stroke on the basis of treatment being started within 3 hours of the start of symptoms. The success of thrombolysis started within 3 hours of stroke onset suggests that a clinically relevant volume of at-risk but salvageable neurons survived until reperfusion started. The earliest that reperfusion can occur is at the beginning of the t-PA infusion, but it is more likely that reperfusion begins during the infusion (i.e., up to 4 hours after symptom onset) or shortly thereafter (perhaps up to 7 hours). Two similar trials have been conducted in Europe—the European Cooperative Acute Stroke Study (ECASS) I3 and ECASS II.4 Both ECASS I and ECASS II allowed treatment with t-PA to be started up to 6 hours after the onset of symptoms. ECASS I used a higher dose of t-PA than the NINDS study (1.1 mg/kg compared with 0.9 mg/kg) but failed to detect any significant advantage for the t-PA group over the placebo group, despite finding effect sizes comparable with those in the NINDS trial. In ECASS I, however, mortality and intracerebral hemorrhages were more common in the t-PA group than in the placebo group. ECASS II used the same dose of t-PA as the NINDS study (0.9 mg/kg). Like ECASS I, ECASS II failed to find any significant benefit of giving patients t-PA.4 Although the study stratified patients according to whether treatment was started less than 3 hours or between 3 and 6 hours after the onset of symptoms, most patients (approximately 80%) began to receive t-PA in the 3- to 6-hour window. A simplistic conclusion from the ECASS studies would be that there are no clinically relevant at-risk neurons that can be rescued if thrombolytic therapy is delayed until more than 3 hours after the onset of symptoms. Using the same assumption as before about the time to reperfusion, this translates to at least 4 to 8 hours after the onset of symptoms. Although the effect of starting thrombolytic treatment more than 6 hours after the onset of symptoms has not been tested, it is unlikely that there are viable at-risk cells after 8 hours if there are none left between 4 and 8 hours.
There are some reasons to conclude that the story is not so simple. The results of a meta-analysis suggest that thrombolytic therapy does have a beneficial effect when started up to 6 hours after the onset of symptoms.5 Approximately half of the data in this meta-analysis came from studies of t-PA. It is therefore possible that the ECASS studies contained a type 2 statistical error, finding no significant effect when in reality the drug did have a beneficial effect, even when started up to 6 hours after symptom onset. It has been pointed out, for example, by Bath,6 that patients in the ECASS studies, particularly in ECASS II, had milder strokes than those in the NINDS study. Patients with mild strokes are more likely to recover without treatment, and therefore the difference between treated and untreated patient groups will be smaller and may not reach statistical significance. The ECASS trials also used different criteria for success than NINDS. In the NINDS study, a global test statistic (the Wald test) was used to assess clinical outcome,7,8 combining scores on the Barthel index,9 the modified Rankin scale,10 the Glasgow outcome scale,11 and the National Institutes of Health Stroke Scale (NIHSS).12 ECASS II, however, used “favorable” outcome rather than “independence” as the primary outcome measure. A post hoc analysis of ECASS II that used independence as the outcome criterion found that t-PA was statistically superior to placebo—treatment with t-PA resulted in an 8.3% increase in the number of patients rated as independent at 90 days. If ECASS II contained a type 2 statistical error, then we may conclude that there may be viable but at-risk neurons as late as 8 hours after the onset of symptoms of ischemic stroke.
Prourokinase The recanalizing efficacy and safety of intra-arterial prourokinase was reported in the PROlyse in Acute Cerebral Thromboembolism (PROACT) I trial.13 This study was not powered to detect any effect of prourokinase on clinical outcome. Consequently, a larger, double-blind, placebo-controlled trial—the PROACT II trial—was performed in patients with acute ischemic stroke.14 In both studies, treatment with prourokinase was started within 6 hours of the onset of symptoms. As primary outcome criteria, the PROACT II investigators used the proportion of patients functioning independently at 90 days as defined by a modified Rankin score of 2 or less. The recruitment criteria were designed to be very demanding in an attempt to reduce the heterogeneity of the sample. Patients were included only with angiographically proven M1 or M2 MCA territory occlusions. This meant that 12,323 patients were screened to identify 180 who were eligible to be randomized to receive either prourokinase
SALVAGING BRAIN TISSUE: THROMBOLYSIS-BASED EVIDENCE
plus intravenous heparin (n ⫽ 121) or intravenous heparin alone (n ⫽ 59). The mean NIHSS score was 17, indicating that these patients had more severe strokes than those in the NINDS, ECASS I, or ECASS II studies. A further refinement of the trial was that patients were stratified according to their NIHSS scores on entry. Overall, the primary analysis indicated that 40% of prourokinase-treated patients were independent at 90 days (modified Rankin score ⱕ 2), compared with only 25% of the control group (P ⫽ .043). Angiographically proven partial or complete recanalization rates 2 hours after the start of the infusion were 66% for the prourokinase group and 18% for the placebo group, although the complete recanalization rates were only 19% and 2%, respectively. The design of the study did not allow the calculation of a mean time to recanalization but did report the median time to starting infusion as being 5.3 hours after stroke onset. Clearly, the recovery rate and recanalization rates (both partial and complete) differed—the rate of recovery was lower than the reperfusion rate. This implies that an appreciable fraction of the patients whose clots were successfully dissolved had no salvageable tissue at the time blood flow was reestablished. It is also clear that many patients did benefit and therefore must have had at-risk tissue that was saved. Taking into account a median time from symptom onset to starting treatment of approximately 5 hours and a 2-hour infusion period, it can be estimated that salvageable tissue remained for up to approximately 7 hours, in at least some patients. Because the figure of 5.3 hours is a median, a number of the patients will have benefited from treatment begun later than this. Nevertheless, using a conservative estimate of 5 hours, the following argument applies: if the ischemic tissue that was saved by thrombolysis was at risk of death from the excitotoxic cascade, the time at which it would have been worth starting neuroprotectant treatment may have extended to up to 7 hours after the onset of symptoms.
Conclusions The evidence from the NINDS study of t-PA indicates that ischemic stroke tissue may be salvageable for up to 5 hours after the onset of symptoms. The PROACT II study of prourokinase suggests a longer time window—a clinically relevant volume of salvageable tissue may remain for at least 7 hours after the onset of symptoms. If reperfusion began after the end of infusion, then there may indeed be viable and salvageable tissue remaining for more than 8 hours after the onset of symptoms in
23
some patients. If the quiescent but salvageable tissue in the ischemic territory is at risk of death because of the excitotoxic cascade, then treatment with thrombolytic plus neuroprotectant should result in the recovery of more tissue than treatment with thrombolytic alone. It is therefore reasonable to conclude from the evidence from studies with thrombolytics that the window for starting a neuroprotectant is at least 7 hours and possibly longer after the onset of stroke symptoms.
References 1. Olney JW. Excitotoxic amino acids and neuropsychiatric disorders. Annu Rev Pharmacol Toxicol 1990;30:47-71. 2. The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. Tissue plasminogen activator for acute ischemic stroke. N Engl J Med 1995; 333:1581-1587. 3. Hacke W, Kaste M, Fieschi C, et al. Intravenous thrombolysis with recombinant tissue plasminogen activator for acute hemispheric stroke. The European Cooperative Acute Stroke Study (ECASS). JAMA 1995;274:1017-1025. 4. Hacke W, Kaste M, Fieschi C, et al. Randomised doubleblind placebo-controlled trial of thrombolytic therapy with intravenous alteplase in acute ischaemic stroke (ECASS II). Second European–Australasian Acute Stroke Study Investigators. Lancet 1998;352:1245-1251. 5. Wardlaw JM, del Zoppo G, Yamaguchi T. Thrombolysis for acute ischaemic stroke. Cochrane Database Syst Rev 2000;2:CD000213. 6. Bath P. Alteplase not yet proven for acute ischaemic stroke [comment]. Lancet 1998;352:1238-1239. 7. Lefkopoulou M, Moore D, Ryan L. The analysis of multiple correlated binary outcomes: Application to rodent teratology experiments. J Am Stat Assoc 1989;84:810-815. 8. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986;42:121130. 9. Mahoney FI, Barthel DW. Functional evaluation: The Barthel Index. Md Med J 1965;14:61-65. 10. van Swieten JC, Koudstaal PJ, Visser MC, et al. Interobserver agreement for the assessment of handicap in stroke patients. Stroke 1988;19:604-607. 11. Teasdale G, Knill-Jones R, van der SJ. Observer variability in assessing impaired consciousness and coma. J Neurol Neurosurg Psychiatry 1978;41:603-610. 12. Lyden P, Brott T, Tilley B, et al. Improved reliability of the NIH Stroke Scale using video training. NINDS TPA Stroke Study Group. Stroke 1994;25:2220-2226. 13. del Zoppo GJ, Higashida RT, Furlan AJ, et al. PROACT: A phase II randomized trial of recombinant pro-urokinase by direct arterial delivery in acute middle cerebral artery stroke. PROACT Investigators. Prolyse in Acute Cerebral Thromboembolism. Stroke 1998;29:4-11. 14. Furlan A, Higashida R, Wechsler L, et al. Intra-arterial prourokinase for acute ischemic stroke. The PROACT II study: a randomized controlled trial. Prolyse in Acute Cerebral Thromboembolism. JAMA 1999;282:2003-2011.