Can the Time Window for Administration of Thrombolytics in Stroke be Increased?

CNS Drugs 2003; 17 (14): 995-1011CURRENT OPINION 1172-7047/03/0014-0995/$30.00/0

© Adis Data Information BV 2003. All rights reserved.

Can the Time Window forAdministration of Thrombolytics inStroke be Increased?Geoffrey A. Donnan,1 David W. Howells,1,2 Romesh Markus,1 Danilo Toni3 andStephen M. Davis4

1 National Stroke Research Institute, Austin & Repatriation Medical Centre, West Heidelberg,Victoria, Australia

2 Department of Medicine, University of Melbourne, Austin & Repatriation Medical Centre,Heidelberg, Victoria, Australia

3 Stroke Unit, Department of Neurological Sciences, University “La Sapienza”, Rome, Italy4 Department of Neurology, Royal Melbourne Hospital, Parkville, Victoria, Australia

Level 1 evidence now shows that thrombolysis in cases of acute ischaemicAbstractstroke is effective if administered within 3 hours of stroke onset. This benefit hasbeen shown to be time dependent and potentially extends beyond 3 hours, withevidence that potentially viable penumbral tissue may be present in a significantproportion of cases well beyond 3–6 hours and, in isolated cases, perhaps up to 48hours. This exposes a ‘stroke recovery gap’, the difference observed between theclinical response to thrombolytic therapy in a given population of patientspresenting with ischaemic stroke and the potential clinical recovery if all of thepenumbra were salvaged under ideal circumstances.

The means of bridging this ‘stroke recovery gap’ using thrombolysis mustinvolve extending the therapeutic time window (i.e. the time between stroke onsetand administration of thrombolytics). Approaches to do this include the use of: (i)improved patient selection with modern neuroimaging techniques, particularlymagnetic resonance imaging using perfusion-weighted image/diffusion-weightedimage mismatch; (ii) newer thrombolytic agents; (iii) lower doses of these agents;(iv) varied methods of administration of thrombolytic therapy including combinedintravenous and intra-arterial approaches; and (v) adjunctive therapies such asneuroprotectants. Should these means of extending the time window for throm-bolysis prove successful, a more widespread use of this form of acute stroketherapy will be possible.

Stroke is the second most common cause of death now changed. Morbidity and/or mortality of patientsglobally and a major cause of disability.[1] Until who present with acute ischaemic stroke is reducedquite recently, there were no management or thera- by the administration of oral aspirin (acetylsalicylicpeutic approaches that had been shown to be effec- acid) within 48 hours[2] and use of the thrombolytictive in improving outcomes. Fortunately, this has agent alteplase (recombinant tissue plasminogen ac-

996 Donnan et al.

tivator; rtPA) within 3 hours[3] of stroke onset. For its efficacy and reduce complication rates. Fortu-nately, these three issues are inter-related, and theyall forms of stroke, management in a stroke unit alsoare the focus of this article. Rather than undertakingimproves outcomes.[4] Of these strategies, throm-a systematic review of the literature, we have pre-bolysis is the most biologically effective. The num-sented our own views and supported these withber of patients needed to treat to benefit one personappropriate evidence.by preventing disability or death associated with

stroke is about 83 for aspirin, 18 for management in1. What is the Time Window fora stroke unit and only 16 for alteplase.[5]

Tissue Salvage?In spite of their biological efficacy, thrombolytic

agents have two major shortcomings. First, the dura-1.1 Tissue Salvage Generallytion of time between stroke onset and administration

of the drug (or the therapeutic ‘time window’) is Tissue salvage relates to the ‘life span’ of theonly 3 hours; second, haemorrhagic complications ischaemic penumbra. The penumbra may be definedmay occur. For the latter, symptomatic intracranial as functionally compromised but structurally intacthaemorrhage rates are about 7%; overall, intracer- tissue usually adjacent to or surrounding a core ofebral haemorrhage rates about double this fig- infarcted tissue.[12] Theoretically, all penumbral tis-ure.[6-11] Currently, because of these limitations, al- sue has the potential to be salvaged at any giventeplase is only administered in about 1–2% of pa- time. In natural history studies, about 50% of pen-tients with ischaemic stroke in the US, although in umbral tissue imaged during the acute phases ofsome specialised centres the rate may be as high as stroke survives and about 50% dies.[13-17] The dura-10%. Clearly, there is a need to extend the time tion of the existence of the penumbra seems to varywindow for thrombolytic therapy, further maximise from case to case in clinical studies.[13] Evidence

from experimental models suggests that the life spanof the penumbra (and time to ischaemic cell death) isdependent upon the degree and duration of the re-duction of cerebral blood flow (CBF).[18]

There are a large number of data available frominvestigators working with animal models of cere-bral ischaemia, where the time window until tissuedeath occurs is reasonably well documented. Theduration is dependent upon the two factors men-tioned above as well as presumed variations in tissuevulnerability in different animal species. The lifespan of the ischaemic penumbra in various animalmodels is less well understood but again is depen-dent upon both the extent and duration of the reduc-tion of CBF.

Memezawa et al.[19] used a rat model of arterialmiddle cerebral artery occlusion to show that fol-lowing 15 minutes of vessel occlusion, selectiveneuronal necrosis in the medial caudoputamen oc-curred, after 30 minutes infarcts extended to thelateral caudoputamen with some involvement of theneocortex, and by 60 minutes cortical infarction wasuniversally present. The infarct size increased pro-

0

1

2

3

4

5

6InfarctPenumbra (18FMISO)

Are

a (m

m2 )

0

1

2

3

4

5

6

8765321 4

Coronal section through infarct

a

b

Fig. 1. Loss of ischaemic penumbra (18F-misonidazole [18FMISO]autoradiography) and growth of infarction over 22 hours after 2hours of transient middle cerebral artery occlusion in rats (n ≥ 6).Areas from coronal sections at eight defined anatomic planes en-compassing the infarct are shown at (a) 1 hour and (b) 22 hourspost-occlusion (courtesy of the National Stroke Research Institute,Heidelberg West, VIC, Australia).

© Adis Data Information BV 2003. All rights reserved. CNS Drugs 2003; 17 (14)

Increasing the Time Window for Thrombolytics in Stroke 997

Heiss et al.[20] were able to furnish evidence ofviable tissue on the border zone of infarcts up to 48hours after stroke in a positron emission tomography(PET) study of 16 patients using CBF, cerebralmetabolic rate of oxygen, cerebral metabolic rate ofglucose and oxygen extraction fraction as markersof tissue viability. Using 18FMISO as a PET markerof hypoxic tissue to identify the penumbra, Read etal.[17] were able to document the presence of signif-icant areas of potentially viable tissue up to 42 hourspost-ischaemic stroke, although this technique as a

Table I. Maximum duration of the ischaemic penumbra in humansafter acute ischaemic stroke as determined using different tech-niques

Reference Technique Penumbral Maximummeasure duration (h)

Heiss et al.[20] PET OEF 48

Read et al.[17] PET 18FMISO 42

Darby et al.[21] MRI PWI/DWI 24mismatch

18FMISO = 18F-misonidazole; DWI = diffusion-weighted imaging;MRI = magnetic resonance imaging; OEF = oxygen extractionfraction; PET = positron emission tomography; PWI = perfusion-weighted imaging.

penumbral marker is still being validated. Longergressively with increasing occlusion time up to duration studies, using magnetic resonance imaging

(MRI) parameters, have been less frequently carriedabout 120–180 minutes, after which the infarct sizeout. Darby et al.[21] showed that perfusion-weighteddid not grow further. This suggests that penumbralimage (PWI)/diffusion-weighted image (DWI) mis-tissue exists in this animal model for only 120–180match as an index of the penumbra exists for at leastminutes. In our hands, using a transient thread oc-24 hours poststroke. The proportion of patients withclusion model in rats and 18F-misonidazoleat least some penumbral tissue were shown in MRI(18FMISO), a hypoxic ligand, as a penumbral mark-and/or 18FMISO PET studies to decline from nearlyer, significant 18FMISO uptake was seen for up to 6100% of patients within the first hours of acutehours, and at least some uptake was still present atischaemic stroke to <40% after 24 hours[22] (figure22 hours (figure 1). After middle cerebral artery2).occlusion with reperfusion in the classical experi-

ments of Jones et al.,[18] thresholds of focal cerebral1.2 Thrombolysis Specificallyischaemia were established in awake monkeys.

When local CBF levels fell below 10cc/100g/minIn animal models of ischaemic stroke, the dura-for 2 hours or below 18cc/100g/min during perma-

tion and time window of benefit of thrombolysis hasnent occlusion, irreversible infarction occurred.been almost randomly rather than systematicallyOne of the great conundrums facing investigatorsexamined. With volume of infarction as the outcome

is how to apply these data from animal models tomeasure, significant reductions have been shown in

humans. There are a number of issues to consider,including the great variety of animal models andtheir relationship to human stroke, lower white mat-ter content of animal brain and other neurochemicaldifferences. Fortunately, in humans, there are a vari-ety of means of imaging the ischaemic penumbra,and a number of these have been used to provide anestimate of the duration of the longest survivingpenumbra.[17,20,21] These are shown in table I.

It is generally accepted that an upper and lowerthreshold of CBF of 20 and 10 mL/100g/min, re-spectively, exists for the ischaemic penumbra duringthe first hours post-acute ischaemic stroke. For thecerebral metabolic rate of oxygen, the threshold isaround 1.40mL/100mg/min.

0

20

40

60

80

100

<12h 12–24h >24h

Time after stroke onset

Pro

port

ion

of p

atie

nts

with

pen

umbr

a (%

)

Fig. 2. 18F-misonidazole positron emission tomography studies ofpatients with acute ischaemic stroke (n = 26), showing that theproportion of patients with at least some penumbra declines overtime (courtesy of the National Stroke Research Institute, HeidelbergWest, VIC, Australia).


998 Donnan et al.

Table II. Animal studies of intravenous (IV) or intra-arterial (IA) administration of thrombolytic agents[23]

Reference Drug Route of Time to first treatment Mean infarct size with treatment,administration after occlusion (min) relative to control size

Andersen et al.[27] Alteplase IV 45 0.65

Carter et al.[28] Alteplase IV 60 0.21

Alteplase IV 120 0.51

Del Zoppo et al.[29] Urokinase IA 180 0.22

Lekieffre et al.[30] Alteplase IV 60 0.45

Overgaard et al.[31] Alteplase IV 120 0.24

Sakurama et al.[32] Alteplase IV 5 0.18



Urokinase IV 5 0.25

Sereghy et al.[33] Alteplase IV 120 0.20

Zhang et al.[34] Tenecteplase-alteplase IA 120 0.60

Tenecteplase-alteplase IA 240 0.82



animals treated with thrombolytics up to 6 hours by functional recovery of the patient). This de-creases steadily as the time since the onset of strokepost-onset of ischaemia compared with controlincreases, to about 48 hours (figure 4). Superim-groups (table II).[23] For humans, as mentioned earli-posed upon this is the actual tissue salvaged usinger, the time window established for efficacy of intra-thrombolysis. As can be seen from the comparison,venous alteplase for acute ischaemic stroke is 3there is considerable room for improvement in: (i)hours, based on the National Institute of Neuro-the amount of tissue that can be salvaged usinglogical Disorders and Stroke (NINDS) trial.[6] Inter-therapeutic agents; and (ii) the time window duringestingly, in the Australian Streptokinase Trialwhich salvage can occur. Narrowing the ‘recovery(ASK), a trend towards efficacy was seen for ther-gap’ or ‘tissue salvage gap’ is the implicit goal of allapy given within 3 hours, although the drug was notresearchers involved in acute stroke therapy trials.effective overall.[24,25] Marler and colleagues[26]

One means of narrowing the stroke recovery gap ishave shown that the benefit of alteplase diminishesto extend the time window for thrombolysis.with time up to 3 hours, with significantly reduced

benefit after this. Based on data from the European2. How Does Thrombolysis Work?Cooperative Acute Stroke Study I (ECASS I)[7] and

the European-Australasian Cooperative Acute Fibrinolysis is triggered by circulating endoge-Stroke Study II (ECASS II),[8] there is no definite nous plasminogen activators (figure 5).[35] Thesebenefit for intravenous alteplase given between 3 are present in very low concentrations (aboutand 6 hours after stroke onset, although a trend 100 000-fold lower than that of plasminogen) andexists. We have conducted a recent meta-analysis of include endogenous plasminogen activators such asall available published data on the use of alteplase in tissue plasminogen activator (tPA) and urokinase-acute ischaemic stroke; the asymptotic effect of time type plasminogen activator (uPA). Recombinanton a steadily reducing benefit can be seen in figure forms of these activators such as alteplase and mu-3. tant forms of the human enzyme tenecteplase[36] can

Based on our understanding of the duration of the be used to trigger the fibrinolytic process. Alterna-ischaemic penumbra in humans, a theoretical model tively, bacterial proteins such as streptokinase orcan be constructed of a maximum amount of tissue staphylokinase may be used.[37] These triggers thenthat can salvaged by an ideal therapy (as represented stimulate the fibrinolytic system by converting plas-



minogen to plasmin. Plasmin then cleaves fibrin into These inhibitors, particularly PAI-1[38] andTAFI,[39] represent important potential targets forits degradation products. When thrombolytic ther-therapy, since their inhibition will enhance the fibri-apy is administered, large amounts of plasminogennolytic effects of exogenously administered plasmi-activators (e.g. tPA, uPA, streptokinase) are used,nogen activators.which result in plasma concentrations that are about

At least one of the plasminogen activators, tPA,1000-fold higher than concentrations of endogenousmay also be potentially neurotoxic.[40] tPA is presentplasminogen activators but still 100-fold lower thanin neurons and microglia and may have other func-concentrations of plasminogen.[35]

tions apart from those in the fibrinolytic pathway.There are two phases in the fibrinolytic process.These include roles in cell migration, astrocyte dif-

In the first phase, plasminogen binds to intact fibrinferentiation and neuron-cell interactions.[41] Tsirka

and initial fibrinolysis occurs. Then, partly degradedet al.[40,42] provided in vitro evidence that alteplase

fibrin exposes new binding sites (carboxy terminal exacerbates excitotoxic neuronal degeneration.residues) for plasminogen. In the second phase, fi- Some support for this was shown in vivo in a mousebrinolysis is enhanced because the local plasmi- stroke model where those animals given alteplasenogen concentration is increased and has higher were shown to have larger areas of focal infarc-reactivity. tion,[43] while tPA-deficient transgenic mice had

The fibrinolytic system is inhibited at at least smaller infarcts than their wild-type litter mates.[43]

However, these findings have not been confirmed inthree important levels (figure 5):[35]

rat models of focal ischaemia where alteplase has1. Plasminogen activator inhibitor 1 (PAI-1) regu-not altered focal infarct size.[44,45] Although thelates plasminogen activators.mechanism of neurotoxicity remains unclear, there

2. Plasmin is inhibited by α2-antiplasmin (α2-AP) is some evidence that tPA may amplify haemo-and to a lesser extent by α2-macroglobulin. globin-induced neurotoxicity in rat neuronal cul-3. Thrombin-activatable fibrinolysis inhibitor tures[46] as well as NMDA- and kainic acid-mediated

neurodegeneration or neuronal death.[40,42,47](TAFI) eliminates carboxy terminal lysine residuesfrom the partially degraded fibrin and hence inhibits Although the findings are of importance concern-the second phase of fibrinolysis. ing possible neurotoxicity of tPA, there is a great

≤3 hours

>3–6 hours

Alteplasen/N

Placebon/N

OR(95% CI fixed)

OR(95% CI fixed)

Alteplase better Placebo better

0.1 0.2 1 5 10

ECASS I 28 / 47 28 / 37 0.47 (0.18, 1.23)ECASS II 47 / 81 48 / 77 0.84 (0.44, 1.58)NINDS 179 / 312 229 / 312 0.49 (0.35, 0.68)

ECASS I 173 / 266 189 / 270 0.80 (0.55, 1.15)ECASS II 197 / 328 200 / 314 0.86 (0.62, 1.18)ATLANTIS 99 / 336 105 / 336 0.92 (0.66, 1.28)

ATLANTIS 8 / 23 16 / 38 0.73 (0.25, 2.14)Total 262 / 463 321 / 464 0.55 (0.42, 0.73)

Total 469 / 930 494 / 920 0.86 (0.71, 1.04)

Fig. 3. Meta-analysis of trials of intravenous alteplase (recombinant tissue plasminogen activator) in patients with acute ischaemic strokestratified by time windows to administration of therapy (0–3 hours or >3–6 hours), in terms of death or dependency as the outcome (courtesyof the National Stroke Research Institute, Heidelberg West, VIC, Australia). ATLANTIS = Alteplase Thrombolysis for Acute Noninterven-tional Therapy in Ischemic Stroke[9]; ECASS I = European Cooperative Acute Stroke Study I[7]; ECASS II = European-AustralasianCooperative Acute Stroke Study II[8]; NINDS = National Institute of Neurological Disorders and Stroke study[6]; OR = odds ratio.


1000 Donnan et al.

Acute ischaemic Stroke (DIAS) study, patients arebeing entered up to 9 hours post-stroke onset.[49]

The EchoPlanar Imaging Thrombolytic Evalua-tion Trial (EPITHET) follows a pilot study in whichpatients who received alteplase 0.9mg/kg werefound to have significantly less infarct growth onMRI (DWI to T2 measure) than historical controlswith similar demographic patterns.[48] This was animportant finding since it was the first therapeuticvalidation of MRI surrogate endpoints in studies ofacute ischaemic stroke. Similar findings were re-ported by Rother et al.[51]

3 6 24 48

Time (h)

Odd

s of

com

plet

e re

cove

ryfo

r de

fined

str

oke

popu

latio

n

Stroke recovery gap

1.0

AlteplasePotential recovery

Fig. 4. The relationship between the odds of recovery to nearnormal functional status (compared with controls) and time of initia-tion of thrombolysis after stroke onset. Although alteplase (recombi-nant tissue plasminogen activator) has improved the odds of recov-ery, there is still a considerable ‘stroke recovery gap’ based on ourunderstanding of the duration of the ischaemic penumbra (courtesyof the National Stroke Research Institute, Heidelberg West, VIC,Australia).

deal of uncertainty as to whether this is clinicallyimportant. However, the issue has stimulated inves-tigators to identify plasminogen activators that havea benign neurotoxic profile and, hence, may havegreater appeal from a marketing perspective.

3. Extending the Time Window

There are a number of potential ways of ex-tending the time window of thrombolysis, includingissues of patient selection, use of newer thromboly-tic agents, varied methods of administration of theseagents and, finally, use of adjunctive therapies.

3.1 Patient Selection

As mentioned in section 1.1, the presence of theischaemic penumbra in individual patients seems tovary considerably with time.[13,17] Hence, to extendthe time window of thrombolysis, selection of pa-tients for thrombolytic therapy who have a largepenumbra at timepoints beyond 3 hours is morelikely to result in clinical improvement. This hypo-thesis is being addressed by a number of investiga-tors using MRI with PWI/DWI mismatch as thepenumbral marker.[48,49] The trials are listed in tableIII. As can be seen, the common time window beingtested is 3–6 hours, but in the Desmoteplase In

tPA uPA

Plasminogen

Plasmin

Plasminogen Plasmin

TAFI

Fibrin FDP

Plasminogen

Intact fibrin Partially degraded fibrin

lys lys lys

lys

lyslys

b

a

PAI-1

α2-AP

α2-M

Fig. 5. (a) The fibrinolytic system: a simplified diagram of the impor-tant elements of thrombolysis showing the points of potential influ-ence of drugs to enhance the process. Plasminogen is activated byconversion to plasmin via tissue plasminogen activator (tPA) andurokinase-type plasminogen activator (uPA). These enzymes areregulated by plasminogen activator inhibitor 1 (PAI-1). Plasmin isresponsible for the degradation of fibrin into fibrin degradation prod-ucts (FDP) and is inhibited by α2-antiplasmin (α2-AP) andα2-macroglobulin (α2-M). Thrombin-activatable fibrinolysis inhibitor(TAFI) is a carboxypeptidase that interferes with the localisation ofplasminogen on the fibrin surface. (b) Increased binding of plasmi-nogen to fibrin: a schematic representation of the weak binding ofplasminogen to intact fibrin and the increased binding to partiallydegraded fibrin. Degradation of intact fibrin by plasmin results in thegeneration of carboxy terminal lysine (lys) residues, which re-present new binding sites for plasminogen (reproduced from Rijkenand Sakharov,[35] with permission from Elsevier).



Table III. Trials of thrombolysis that assessed early perfusion-weighted imaging (PWI) and diffusion-weighted imaging (DWI) parameters todetermine ideal patterns to prolong the therapeutic window

Trial Drug Dose and route of Penumbral Study time Sample size Fundingadministration assessment window (h)

DEFUSE[50] Alteplase 0.9 mg/kg IV MRI PWI/DWI 3–6 80 National Institutes ofmismatch Health, USA

DIAS[49] Desmoteplase Escalating IV dose MRI PWI/DWI 3–9 Tiered phase II; PAION Company,mismatch 690 phase III Germany

EPITHET[48] Alteplase 0.9 mg/kg IV MRI PWI/DWI 3–6 120 National Health andmismatch Medical Research

Council, Australia

DEFUSE = Diffusion-weighted imaging Evaluation For Understanding Stroke Evolution; DIAS = Desmoteplase In Acute ischaemic Stroke;EPITHET = EchoPlanar Imaging Thrombolytic Evaluation Trial; IV = intravenous; MRI = magnetic resonance imaging.

Another issue of importance relating to these imaging selection criteria to 3–5 hours (ECASSIII[52]) and up to 6 hours (The Third Internationaltrials is the inclusion of patients both with andStroke Trial [IST-3][53]).without PWI/DWI mismatch (i.e. those with and

without penumbra, respectively). In EPITHET, pa-tients are randomised between 3 and 6 hours post- 3.2 Newer Thrombolytic Agentsstroke onset, with both mismatch and nonmismatchpatients included.[48] Hence, if there is a differential Because of the issues of haemorrhagic complica-

tions and possible neurotoxicity, there is a constantof MRI surrogate outcomes between these two MRIdrive to develop agents with better adverse effectsubgroups, the capacity for it to be detected exists.profiles. In general, the agents discussed in sectionsAlthough it seems logical that those with greater3.2.1–3.2.5 have at least one of two advantages overmismatch are likely to respond better to thromboly-alteplase: lower toxicity because of higher fibrintic therapy, this has not as yet been proven. Con-specificity or a reported absence of neurotoxicity.versely, in the DIAS trial, only patients with mis-Clearly, should this be the case, the time window formatch (at least 20%) are included prior to randomis-administration of thrombolysis may be extended.ation. Currently, DIAS patients are being studied inAlthough the discussion is not exhaustive, some ofa phase II design with tiered doses of the bat serinethe more important agents are covered.protease desmoteplase (15 patients are receiving

desmoteplase and four placebo in each tier) in an3.2.1 Staphylokinaseescalating-dose, weight-adjusted design. In theStaphylokinase is highly fibrin selective, the mo-phase III study, patients will be equally randomised

lecular mechanisms of which have been recentlyto desmoteplase (n = 345) or placebo (n = 345).reviewed.[54] There are two important mechanismsAlthough patients entering the Diffusion-weightedto consider. First, staphylokinase has only a lowimaging Evaluation For Understanding Strokeaffinity for plasminogen in the circulation and plas-Evolution (DEFUSE) trial[50] are not randomised,minogen bound to intact fibrin but a high affinity forboth nonmismatch and mismatch patients are in-plasminogen bound to partially degraded fibrin and

cluded so that important information about thefor plasmin. This results in an accumulation of

differential effect on MRI surrogate outcomes men-staphylokinase on the surface of a lysing clot.[55]

tioned above will also be gleaned.Second, the plasmin-staphylokinase complex is rap-

For all of these trials, hypotheses concerning idly inhibited by α2-AP in the absence of fibrin (inMRI characteristics that may predict haemorrhagic the circulation) and only slowly in the presence oftransformation and haematoma formation may also fibrin (on the clot surface). Staphylokinase, as yet,be tested. Other investigators have also taken the has not been studied extensively in ischaemicapproach of extending the time window without stroke.[37,56]


1002 Donnan et al.

3.2.2 Tenecteplase desmoteplase is by direct activation of plasminogenLike other newer thrombolytic agents, tenecte- to plasmin (figure 5a).[49] Desmoteplase appears to

plase is more fibrin-specific than alteplase be more fibrin specific than alteplase and therefore(14-fold).[57] Interestingly, it may also have a rela- may have a lower haemorrhagic transformation pro-tive resistance to PAI-1.[36] It also has a longer file. Further, desmoteplase did not appear to havebiological half-life such that bolus administration is the neurotoxicity of tPA when the effects of kainicpossible. Its high fibrin specificity allows tenecte- acid striatal toxicity were studied.[47] A phase IIAplase to be active at the site of high fibrin concentra- trial in patients with acute myocardial infarctiontions, such as within recent clots, rather than system- (DEEP) has recently been conducted. For acute is-ically. Tenecteplase has already been studied in chaemic stroke, the MRI-based DIAS trial (see sec-trials of ischaemic heart disease[58-60] and, in general, tion 3.1) is currently underway, and a similar trial,does not appear to be associated with any increase in the Dose Escalation Study of Desmoteplase in Acutethe incidence of cerebral haemorrhage compared Ischemic Stroke (DEDAS), is planned for the US.with alteplase. Another theoretical advantage of ten-ecteplase is the lack of a ‘plasminogen steal’ ef- 3.2.4 Pro-urokinasefect[59] (a paradoxical decrease in the efficacy of

Pro-urokinase is a glycosylated 411-amino-acidalteplase may be seen when plasminogen leechessingle-chain proenzyme precursor of urokinase. It isfrom the clot in the presence of high alteplase con-derived from transfected murine SP2/0 hybridomacentrations).[61]

cells.[64] Single-chain r-pro-urokinase is activated toThere have been a number of promising studiestwo-chain urokinase at the thrombus surface by fi-in animal models suggesting that tenecteplase maybrin-associated plasmin.[65] Interestingly, the throm-be a useful alternative to alteplase. In a rabbit em-bolytic effect of pro-urokinase is augmented by he-bolic stroke model, tenecteplase-treated animals hadparin, possibly through thrombin neutralisation orfewer associated cerebral haemorrhages than con-by stimulating tPA release from the endotheli-trols,[57] although in a comparative study betweenum.[66,67]

alteplase and tenecteplase in a similar model[62]

After a pilot study (Pro-urokinase in Acute Cere-there was no significant difference between the twobral Thromboembolism [PROACT] I), PROACT II(similar cerebral haemorrhage rates). In a rat embol-was conducted in 54 centres in the US and Cana-ic model of ischaemic stroke, a tenecteplase-al-da.[10] 180 patients with acute ischaemic stroke of <6teplase combination was shown to be effective inhours’ duration caused by angiographically provenreducing neurological deficits and ischaemic vol-occlusion of the middle cerebral artery withoutume without increasing haemorrhagic transforma-haemorrhage or major early infarction signs ontion when administered up to 4 hours post-onset ofcomputed tomography were randomised to receiveischaemia.[63]

intra-arterial pro-urokinase plus heparin (n = 121) orFor the above reasons, tenecteplase-alteplaseheparin alone (n = 59). For the primary analysis,does have some theoretical advantages over al-40% of pro-urokinase patients and 25% of controlteplase and, given its demonstrated safety in animalpatients had a modified Rankin score of ≤2 (p =models, clinical trials in ischaemic stroke are war-0.04). The recanalisation rate was 66% for the pro-ranted.urokinase group and 18% for the control group (p <0.01). Ten percent of patients had symptomatic in-3.2.3 Desmoteplasetracerebral haemorrhage.[10] Unfortunately, in spiteDesmoteplase is a serine protease that has beenof the positive outcome of the study, the US FDAgenerated by recombinant biotechnology in Chinesedid not approve pro-urokinase for standard therapyhamster ovary cells. The naturally occurring com-in acute ischaemic stroke. A further study is plannedpound was discovered to exist in the saliva of the bat(PROACT III) but has not, as yet, commenced.Desmodus rotundus. The mechanism of action of



3.2.5 Ancrod of 0.8 mg/kg (10% administered as a bolus, with theAncrod can be loosely termed a thrombolytic remainder over 90 minutes) in an open observational

agent in that it produces rapid defibrinogenation by study. The time window adopted was 7 hours, thesplitting fibrinopeptide A from fibrinogen.[68,69] This longest yet studied systematically. This study alsocauses anticoagulation by depleting the substrate involved anticoagulation with heparin or nadro-needed for thrombus formation. Blood viscosity is parin. Intravenous heparin (to achieve an activatedalso decreased, and the products of defibrinogena- partial thromboplastin time of 1.5 × baseline value)tion may also enhance local clot-specific throm- was given immediately after alteplase (n = 31) orbolysis by stimulating endogenous plasminogen ac- within 24 hours (n = 41), while subcutaneous nadro-tivators.[70] The removal of fibrinogen also has an parin was given immediately after alteplase (n = 9;antiaggregant effect on platelets, because it is a dose of 3075IU anti-XaIU). At 3 months, 45% ofligand for aggregation at the IIb/IIIa receptor patients had a good outcome (modified Rankin scoresites.[71] 0–1), and the overall symptomatic haemorrhage rate

After two small, double-blind, randomised, pla- was 7%.cebo-controlled trials of ancrod with time windows Since these safety and outcome profiles are simi-of 6 hours suggested a possible benefit on clinical lar to those in the NINDS study,[6] even in theoutcome measures (Scandinavian Stroke Scale), the presence of heparinisation, the use of longer timeStroke Treatment with Ancrod Trial (STAT) was windows with lower doses of thrombolytic agentsconducted in the US.[72] In this trial, 500 patients may be worthwhile.were randomised to receive either ancrod or placebowithin 3 hours of ischaemic stroke onset. A signif- 3.4 Varied Methods of Administrationicant outcome benefit was seen for ancrod (favoura-

The great advantage of intravenous administra-ble outcome of Barthel index ≥95: 42.2% in thetion of alteplase is that it may be given quickly in aancrod group vs 34.4% in the placebo group; p =variety of clinical settings. However, its major dis-0.04). Symptomatic bleeding occurred in 5.2% ofadvantage is that the concentration of drug at the sitepatients treated with ancrod and 2.0% in the placeboof vessel occlusion may not be great enough to causegroup. A better outcome was achieved with lowerlysis in every case. In these circumstances, it seemsfibrinogen concentrations, and symptomatic in-logical that intra-arterial administration may betracerebral haemorrhage was more likely to occurmore appropriate, since higher local but lower sys-under these circumstances.temic concentrations of the thrombolytic agent mayA parallel European trial with ancrod using abe given and hence lower systemic complication6-hour time window was also commenced, but therates incurred. Further, mechanical disruption of thestudy was prematurely terminated because of aclot also may be possible during the angiographicfailed futility assessment at a pre-planned interimprocedure. The paradigm for the intra-arterial ap-analysis.[73] Further, the 90-day mortality was higherproach was shown in the PROACT II trial discussedin the ancrod group than the placebo group.in section 3.2.4.[10]

It is unclear whether this agent will be developedThe combined intravenous/intra-arterial ap-further using the shorter 3-hour time window used in

proach has been adopted by a number of investiga-the US trial.tors, as summarised in table IV. It was first studiedby Freitag et al.,[76] who used combined intravenous3.3 Lower Doses of Thrombolytic Agentsalteplase with intra-arterial plasmin. The combina-

The use of lower doses of thrombolytic agents is tion produced higher recanalisation rates than hada logical approach to reduce adverse effects. It has been reported previously with either approach alone.been adopted by Trouillas and colleagues,[74,75] who In the Emergency Management of Stroke (EMS)treated 100 patients with a reduced dose of alteplase Bridging Trial,[77] combination intravenous/intra-ar-


1004 Donnan et al.

terial alteplase versus intra-arterial alteplase alone showing at least an additive effect.[27,30,31,33,88-94]

was used within 3 hours of stroke onset. The advan- These are summarised in table V. Important pointstage of this trial was its double-blind, randomised, to note are that the experimental models have allplacebo-controlled design. Again, combined treat- involved elements of reperfusion, including the ratment resulted in higher recanalisation rates (53% vs thromboembolic model, which is perhaps closest to28%) with an acceptable safety and outcome profile; the human paradigm. However, because of the vari-there was no difference in clinical outcome between able extent of infarction in this model, somewhatthe two groups although there were more deaths in larger sample sizes are required. This may, in part,the intravenous/intra-arterial group (see table IV). be an explanation for the negative findings of Over-Other smaller open studies[78,79] also showed superi- gaard et al.,[31] who used a combination ofor recanalisation rates for combined intravenous/ 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinox-intra-arterial administration (about 70–80%) com- aline (NBQX) with alteplase and showed that al-pared with previously reported rates with intrave- though both drugs individually reduced infarct size,nous therapy alone. Two studies[80,81] have used a there was no significant additive effect with thecombined intravenous/intra-arterial approach to- combination. Limited sample sizes may also ac-gether with selection of patients based on MRI PWI/ count for negative findings with other neuropro-DWI criteria. Again, recanalisation rates were great- tectants.[93,94] This issue needs to be explored furtherer (≥80%) than those previously reported with intra- in larger animal models, using differing time win-venous therapy alone. Finally, a reverse approach dows to therapy to get a better understanding of thewith intra-arterial therapy preceding the intravenous conditions under which additive effects may occur.contribution has been studied openly, the rationale In clinical studies, far fewer data are availablebeing that the early intra-arterial approach would (table VI).[95-97] Unfortunately, sample sizes have allmaximise the probability of recanalisation while been too small to assess clinical outcomes but, im-continuous intravenous infusion would increase the portantly, combination therapies have raised noefficacy of therapy.[82] For all of these combined safety issues. In the largest study, the Acute Strokeapproach studies, sample sizes were too small to Therapy by Inhibition of Neutrophils (ASTIN) trial,sensibly interpret outcome data apart from recanal- approximately 20% of patients with acute ischaemicisation rates. Pleasingly, haemorrhagic complication stroke treated with neutrophil inhibitory factorrates above those seen in the NINDS study were not (NIF) within 6 hours of ischaemic stroke onset re-observed.[26] ceived alteplase.[97] There was no reported inter-

action of effect between alteplase and NIF and,3.5 Adjunctive Therapies equally importantly, no safety issues were raised.

3.5.2 Drugs Enhancing The Thrombolytic Process3.5.1 NeuroprotectantsThe principal of combination neuroprotectant Plasminogen Activator Inhibitor-1 Inhibitors

plus thrombolysis therapy is 2-fold. First, a The PAI-1 inhibitor is the most important inhibi-neuroprotectant administered early after stroke on- tor of tPA (figure 5a). Although it is present inset may ‘freeze’ the penumbra, slowing the cascade relatively low concentrations in plasma, its localof neurochemical events until reperfusion can be concentration in platelet-rich thrombi is relativelyachieved using thrombolysis. Second, the neuropro- high because of its presence in platelets. Hence, it istectant may act as a free radical scavenger, limiting likely to be extremely important during thromboly-reperfusion injury to penumbral tissue once throm- tic therapy. Of interest is that tenecteplase-alteplasebolysis is achieved.[84-87] has a relative resistance to PAI-1.[36] Lindgren et

There is a considerable number of data from al.[98] studied 135 stroke patients and 77 controlsexperimental studies in animal models to support within 7 days after stroke onset and again 2–4 yearsboth of these views, with a number of combinations later (a subset of 32 patients and 18 controls). The


Increasing the Tim

e Wind

ow for T

hrombolytics in Stroke

1005

© A

dis D

ata

Info

rma

tion

BV 2003. A

ll righ

ts rese

rved

.C

NS D

rug

s 2003; 17 (14)

Table IV. Combined intravenous (IV)/intra-arterial (IA) thrombolysis

Reference Study design Sample Dose of IV IV RCN (%) Dose of IV/IA IV/IA RCN (%) Commentsize alteplase (mg/ alteplase (mg/kg)

kg) [hours afterstroke onset]

Ernst et al.[78] RT, open 20 0.6 up to 60mg 0.3 up to 24mg 69 Rankin score 0–2, 63%

Freitag et al.[76] NR, open 20 10mg (total) a Lysine-plasminogen as adjunct

Hill et al.[81] NR, open 7 0.9 0.9 up to 20mg 86 PWI/DWI mismatch for selection; IA given

before IV

IMS[83] NR 80 0.6 Up to 22mg (2h) 57 Rankin 0–1, 30%; average alteplase dose

59mg (NINDS 69mg); ICH 6%

Keris et al.[82] NR, open 45 0.6 25mg (total) 50 Reverse approach IA/IV

Lewandowski et al.[77] RCT, IV/IA 34 0.6 (maximum 28 (placebo) 20mg over 2h 53 No difference in clinical outcome; more

(EMS Bridging Trial) alteplase vs IV 60mg) [≤3h] deaths in the active treatment group; ICH

placebo/IA not different

alteplase

Suarez et al.[80] NR, open 45 0.6 [≤3h] b 75 Patients selected for IA based on PWI/DWI

mismatch and MRA

Zaidat et al.[79] RT, open 18 0.6 up to 30mg 80 Patients studied with carotid occlusion

a IA 2500 activated lysine-plasminogen.

b Alteplase, urokinase based on MRA, MRI PWI/DWI.

DWI = diffusion-weighted imaging; EMS = Emergency Management of Stroke; ICH = intracerebral haemorrhage; IMS = Intravenous Management of Stroke study; MRA = magnetic

resonance angiography; MRI = magnetic resonance imaging; NINDS = National Institutes of Neurological Disorders and Stroke; NR = nonrandomised; PWI = perfusion-weighted

imaging; RCN = recanalisation; RCT = randomised controlled trial; RT = retrospective.

1006 Donnan et al.

tPA antigen concentration was found to be an inde-pendent discriminator between ischaemic stroke pa-tients and controls, as were PAI-1 concentrations (p= 0.05). At re-examination 2–4 years later, neithertPA nor PAI-1 concentrations had changed signifi-cantly from baseline values. The high PAI-1 con-centrations infer increased fibrinolytic inhibition inacute ischaemic stroke and may represent a thera-peutic target to enhance thrombolysis. In other stud-ies, normal[99,100] concentrations of PAI-1 have beendocumented in the acute phases of stroke. To date,no experimental models or clinical stroke studieshave been conducted.

Direct Thrombin Inhibitors

Direct thrombin inhibitors are, more accurately,anticoagulants. Thrombin is central in the thrombot-ic process in that it is a potent platelet agonist andrecruits additional platelets after being exposed totissue factor. Thrombin also converts fibrinogen tofibrin, which then stabilises platelet aggregates. Di-rect thrombin inhibitors were developed to helpovercome the problem of the heparin/antithrombincomplex to inactivate bound thrombin satisfactori-ly.[101] Direct thrombin inhibitors have been exten-sively studied in myocardial ischaemia, and a meta-analysis was performed recently.[102] Based on thismeta-analysis, direct thrombin inhibitors seem to besuperior to heparin for the prevention of death ormyocardial infarction in patients with acute corona-ry syndromes. Compared with heparin, there was anincreased risk of major bleeding with the directthrombin inhibitor hirudin but a reduction withbivalirudin.

Little work with this class of drugs has been donein acute ischaemic stroke. Mikulski et al.[103] showedthat oral or intravenous direct thrombin inhibitorsreduced the size of photochemically induced corti-cal infarction in rats. Given the failure of heparin toimprove outcomes in acute ischaemic stroke, it isuncertain whether direct thrombin inhibitors will bestudied extensively in clinical stroke trials.

Thrombin-Activatable Fibrinolysis Inhibitors

TAFI has been only recently isolated from plas-ma.[39] It represents a link between coagulation andfibrinolysis because its pro-enzyme is activated by


Tab

le V

. E

xper

imen

tal s

tudi

es u

sing

ani

mal

mod

els

of in

tera

ctio

n be

twee

n th

rom

boly

tic a

nd n

euro

prot

ectiv

e ag

ents

Ref

eren

ceN

euro

prot

ecta

ntM

echa

nism

of

actio

nT

hrom

boly

ticE

xper

imen

tal m

odel

Out

com

eag

ent

And

erse

n et

al.[2

7]C

itico

line

Mem

bran

e re

pair,

red

uces

FF

aA

ltepl

ase

Rat

thr

ombo

embo

licA

dditi

vere

leas

e

Bow

es e

t al

.[90]

Mon

oclo

nal a

ntib

odie

s ag

ains

tA

nti-i

nfla

mm

ator

yA

ltepl

ase

Rab

bit

Add

itive

leuc

ocyt

e ad

hesi

on m

olec

ules

Leki

effr

e et

al.[3

0]E

lipro

dil

NM

DA

rec

epto

r an

tago

nist

(po

ly?)

Alte

plas

eR

at t

hrom

boem

bolic

Add

itive

VS

CC

ant

agon

ist

Oro

zco

et a

l.[94]

21 a

min

o st

eroi

d (U

7+00

6F)

Fre

e ra

dica

l sca

veng

erA

ltepl

ase

Rab

bit

No

addi

tive

effe

ct

Ove

rgaa

rd e

t al

.[31]

NB

QX

AM

PA

rec

epto

r an

tago

nist

Alte

plas

eR

at t

hrom

boem

bolic

No

addi

tive

effe

ct

San

chez

et

al.[9

3]U

-743

89-G

Fre

e ra

dica

l sca

veng

erA

ltepl

ase

Rat

thr

ombo

embo

licN

o ad

ditiv

e ef

fect

Ser

eghy

et

al.[3

3]D

izoc

ilpin

eN

MD

A r

ecep

tor

anta

goni

stA

ltepl

ase

Rat

thr

ombo

embo

licA

dditi

ve

Shu

aib

et a

l.[89]

Citi

colin

eM

embr

ane

repa

ir, r

educ

es F

Fa

Uro

kina

seR

at t

hrom

boem

bolic

Add

itive

rele

ase

Yan

g et

al.[8

8]T

opira

mat

eE

nhan

ces

GA

BA

, in

hibi

ts g

luta

mat

eU

roki

nase

Rat

Add

itive

rele

ase

Zha

ng e

t al

.[91]

Mon

oclo

nal a

ntib

odie

s ag

ains

tA

nti-i

nfla

mm

ator

yA

ltepl

ase

Rat

Add

itive

leuc

ocyt

e ad

hesi

on m

olec

ules

Zha

ng e

t al

.[92]

Pro

teas

ome

inhi

bito

r (P

S-5

19)

Ant

i-inf

lam

mat

ory

Alte

plas

eR

at t

hrom

boem

bolic

Add

itive

FF

a =

fre

e fa

tty a

cid;

NB

QX

= 2

,3-d

ihyd

roxy

-6-n

itro-

7-su

lfam

oyl-b

enzo

[f]qu

inox

alin

e; V

SC

C =

vol

tage

-sen

sitiv

e ca

lciu

m c

hann

els.


thrombin during blood clotting. Potato carboxy pep-tidase inhibitor (PCPI) can inhibit TAFI. The prima-ry mechanism of action of PCPI relates to thecarboxy peptidase B activity of TAFI, whichremoves carboxy terminal lysine residues from de-grading fibrin and, possibly, cellular binding sites ofplasminogen.[104] There is some evidence that exog-enous lysis during thrombolytic therapy might beinfluenced by the TAFI system. For example, anti-bodies against TAFI increased endogenous lysis2-fold in a rabbit jugular vein thrombolysismodel.[105] Further, models of thrombolysis in rab-bits show that the administration of PCPI togetherwith alteplase significantly improves alteplase-in-duced thrombolysis.[106-108]

3.5.3 Drugs ReducingHaemorrhagic ComplicationsUnder conditions of brain injury such as ischae-

mia, haemorrhage or trauma, it has been shown thata class of zinc-dependent matrix metalloproteinases(MMPs) are upregulated.[109,110] In their role as ex-tracellular proteases, it has been shown that MMPshave the ability to degrade the extracellular matrix.After reperfusion (e.g. with alteplase), uncontrolledMMP activation may result in degradation of pro-teins such as collagen and laminin within the basillamina. This phenomenon seems to be associatedwith haemorrhagic transformation, probably be-cause of disruption of these membrane proteinswithin vascular structures, thus leading to diapedesisand haemorrhage.[109,110] There is a relationship be-tween free radical oxidative damage and reperfusioninjury. This may be via the link between free radicalinjury and MMP upregulation, which has been re-cently demonstrated. Interestingly, alteplase mayalso be linked biochemically with MMP proteolysis,with plasmin acting as an intermediary step andupstream activator of the MMP cascade.[111]

There have been several experimental studies inwhich MMP inhibitors have been shown to reducehaemorrhagic transformation associated with al-teplase.[112,113] Similarly, the spin-trap agent α-phen-yl-N-t-butylnitrone was found to reduce the amountof haemorrhagic transformation, in a focal rabbitmodel of cerebral ischaemia.[113] This suggests that


Tab

le V

I. C

linic

al s

tudi

es o

f th

e tr

eatm

ent

of a

cute

isch

aem

ic s

trok

e w

ith c

ombi

natio

n th

erap

y

Ref

eren

ceN

euro

prot

ecta

ntM

echa

nism

Thr

ombo

lytic

age

ntS

tudy

des

ign

Sam

ple

Tim

e w

indo

w t

oO

utco

me

of a

ctio

nsi

zetr

eatm

ent

(h)

Gro

tta[9

5]IV

lube

luzo

le 7

.5m

g ov

erM

ultip

leIV

alte

plas

e 0.

9 m

g/kg

RC

T90

≤3N

o sa

fety

issu

es

1 ho

ur,

then

con

tinuo

us

5-da

y in

fusi

on o

f

10 m

g/da

y

Lyde

n et

al.[9

6]IV

clo

met

hiaz

ole

Mul

tiple

IV a

ltepl

ase

0.9

mg/

kg,

RC

T10

1≤3

No

safe

ty is

sues

68 m

g/kg

max

imum

90m

g

Hac

ke e

t al

.[97]

IV N

IF,

esca

latin

g do

seN

eutr

ophi

lIV

alte

plas

e 0.

9 m

g/kg

Ada

ptiv

e R

CT

966

<6

No

safe

ty is

sues

(AS

TIN

stu

dy)

inhi

bitio

n

AS

TIN

= A

cute

Str

oke

The

rapy

by

Inhi

bitio

n of

Neu

trop

hils

; IV

= in

trav

enou

s; N

IF =

neu

trop

hil i

nhib

itory

fac

tor;

RC

T =

ran

dom

ised

con

trol

led

tria

l.

1008 Donnan et al.

3. The National Institute of Neurological Disorders and Stroke rt-free radical scavenging per se may also be useful inPA Stroke Study Group. Tissue plasminogen activator for

minimising haemorrhagic complications of alteplase acute ischemic stroke. N Engl J Med 1995; 333 (24): 1581-74. Langhorne P, Williams BO, Gilchrist W, et al. Do stroke unitsadministration.

save lives? Lancet 1993; 342 (8868): 395-8Research is ongoing to identify patients who are5. Hankey GJ, Warlow CP. Treatment and secondary prevention of

more likely to develop haemorrhagic complications stroke: evidence, costs, and effects on individuals and popula-tions. Lancet 1999; 354 (9188): 1457-63by re-evaluating existing trial data, accruing new

6. Kwiatkowski TG, Libman RB, Frankel M, et al. Effects ofinformation about thrombolysis with longer time tissue plasminogen activator for acute ischemic stroke at onewindows (IST-3) and using neuroimaging tech- year: National Institute of Neurological Disorders and Stroke

Recombinant Tissue Plasminogen Activator Stroke Studyniques such as MRI.Group. N Engl J Med 1999; 340 (23): 1781-7

7. Hacke W, Kaste M, Fieschi C, et al. Intravenous thrombolysis4. Conclusions and Future Directions with recombinant tissue plasminogen activator for acute hemi-

spheric stroke: the ECASS Study Group. JAMA 1995; 274:We are still in the very early phases of under- 1017-25

8. Hacke W, Kaste M, Fieschi C, et al. Randomised double-blindstanding how to use thrombolysis effectively in pa-placebo-controlled trial of thrombolytic therapy with intrave-

tients with acute ischaemic stroke, as evidenced by nous alteplase in acute ischaemic stroke (ECASS II): SecondEuropean-Australasian Acute Stroke Study Investigators. Lan-the wide ‘tissue salvage’ or ‘stroke recovery’ gap.cet 1998; 352 (9136): 1245-51Given that we know that thrombolysis works (intra-

9. Clark WM, Wissman S, Albers GW, et al. Recombinant tissue-venous alteplase, intra-arterial urokinase), a top pri- type plasminogen activator (alteplase) for ischemic stroke 3 to

5 hours after symptom onset. The ATLANTIS Study: a ran-ority in future and ongoing research should be todomized controlled trial: Alteplase Thrombolysis for Acuteextend the time window beyond the current 3-hour Noninterventional Therapy in Ischemic Stroke. JAMA 1999;282 (21): 2019-26time limit by using existing drugs (particularly al-

10. Furlan A, Higashida R, Wechsler L, et al. Intra-arterialteplase) in varying doses and by varying routes ofprourokinase for acute ischemic stroke: the PROACT II study.

administration. As our knowledge about the is- A randomized controlled trial: prolyse in acute cerebral throm-boembolism. JAMA 1999; 282 (21): 2003-11chaemic penumbra and, more importantly, our abili-

11. Wardlaw J, del Zoppo G, Yamaguchi T. Thrombolysis for acutety to assess it rapidly in vivo increases, tools such asischaemic stroke (Cochrane Review). The Cochrane Database

MRI PWI/DWI will become more widely used to of Systematic Reviews. Available in The Cochrane Library[database on disk and CD ROM]. Updated quarterly. Theidentify patients who are more likely to benefit fromCochrane Collaboration; issue 3. Oxford: Update Software,

thrombolysis. The next priority is to augment the 1999effect of thrombolysis or prolong the time window 12. Astrup J, Symon L, Siesjo BK. Thresholds in cerebral ischemia:

the ischemic penumbra. Stroke 1981; 17: 723-5available for its use by administering adjunctive13. Baron J, von Kummer R, del Zoppo G. Treatment of acute

agents. The challenge will then be to translate exper- ischemic stroke: challenging the concept of a rigid and univer-sal time window. Stroke 1995; 26: 2219-21imental data into the clinical sphere. This task, so

14. Marchal G, Beaudouin V, Rioux P, et al. Prolonged persistencefar, has proven to be the most difficult.of substantial volumes of potentially viable brain tissue afterstroke: a correlative PET-CT study with voxel-based data

Acknowledgements analysis. Stroke 1996; 27: 599-60615. Heiss WD, Thiel A, Grond M, et al. Which targets are relevant

No sources of funding were used to prepare this manu- for therapy of acute ischemic stroke? Stroke 1999; 30 (7):script. The authors have received financial support from a 1486-9large number of pharmaceutical companies, but have no 16. Heiss WD, Kracht LW, Thiel A, et al. Penumbral probability

thresholds of cortical flumazenil binding and blood flow pre-conflicts of interest that are directly relevant to the contents ofdicting tissue outcome in patients with cerebral ischaemia.this article.Brain 2001; 124 Pt 1: 20-9

17. Read SJ, Hirano T, Abbott DF, et al. The fate of hypoxic tissueon 18F-fluoromisonidazole positron emission tomographyReferencesafter ischemic stroke. Ann Neurol 2000; 48 (2): 228-351. Murray CJ, Lopez AD. Global mortality, disability, and the

18. Jones TH, Morawetz RB, Crowell RM, et al. Thresholds ofcontribution of risk factors: Global Burden of Disease Study.focal cerebral ischemia in awake monkeys. J Neurosurg 1981;Lancet 1997; 349 (9063): 1436-4254: 773-822. International Stroke Trial Collaborative Group. The Internation-

al Stroke Trial (IST): a randomised trial of aspirin, subcutane- 19. Memezawa H, Smith ML, Siesjo BK. Penumbral tissues sal-ous heparin, both, or neither among 19435 patients with acute vaged by reperfusion following middle cerebral artery occlu-ischaemic stroke. Lancet 1997; 349 (9065): 1569-81 sion in rats. Stroke 1992; 23 (4): 552-9



20. Heiss WD, Huber M, Fink GR, et al. Progressive derangement 39. Bajzar L, Manuel R, Nesheim ME. Purification and characteri-of periinfarct viable tissue in ischemic stroke. J Cereb Blood zation of TAFI, a thrombin-activable fibrinolysis inhibitor. JFlow Metab 1992; 12 (2): 193-203 Biol Chem 1995; 270 (24): 14477-84

40. Tsirka SE, Rogove AD, Strickland S. Neuronal cell death and21. Darby DG, Parsons MW, Barber PA, et al. Significance of acutetPA. Nature 1996; 384 (6605): 123-4multiple brain infarction on diffusion-weighted imaging.

Stroke 2000; 31 (9): 2270-1 41. Pittman RN, Di Benedetto AJ. PC12 cells overexpressing tissueplasminogen activator regenerate neurites to a greater extent22. Read SJ. Positron emission tomography studies of acute strokeand migrate faster than control cells in complex extracellularusing 18F-fluoromisonidazole [dissertation]. Heidelberg:matrix. J Neurochem 1995; 64 (2): 566-75University of Melbourne, 1999

42. Tsirka SE, Rogove AD, Bugge TH, et al. An extracellular23. Jonas S, Aiyagari V, Vieira D, et al. The failure of neuronalproteolytic cascade promotes neuronal degeneration in theprotective agents versus the success of thrombolysis in themouse hippocampus. J Neurosci 1997; 17 (2): 543-52treatment of ischemic stroke: the predictive value of animal

43. Wang YF, Tsirka SE, Strickland S, et al. Tissue plasminogenmodels. Ann N Y Acad Sci 2001; 939: 257-67activator (tPA) increases neuronal damage after focal cerebral24. Donnan GA, Davis SM, Chambers BR, et al. Streptokinase forischemia in wild-type and tPA-deficient mice. Nat Med 1998;acute ischemic stroke with relationship to time of administra-4 (2): 228-31tion: Australian Streptokinase (ASK) Trial Study Group.

44. Meng W, Wang X, Asahi M, et al. Effects of tissue typeJAMA 1996; 276 (12): 961-6plasminogen activator in embolic versus mechanical models of25. Cornu C, Boutitie F, Candelise L, et al. Streptokinase in acutefocal cerebral ischemia in rats. J Cereb Blood Flow Metabischemic stroke: an individual patient data meta-analysis: the1999; 19 (12): 1316-21Thrombolysis in Acute Stroke Pooling Project. Stroke 2000;

45. Klein GM, Li H, Sun P, et al. Tissue plasminogen activator does31 (7): 1555-60not increase neuronal damage in rat models of global and focal26. Marler JR, Tilley BC, Lu M, et al. Early stroke treatmentischemia. Neurology 1999; 52 (7): 1381-4associated with better outcome: the NINDS rt-PA stroke study.

46. Wang X, Asahi M, Lo EH. Tissue type plasminogen activatorNeurology 2000; 55 (11): 1649-55amplifies hemoglobin-induced neurotoxicity in rat neuronal27. Andersen M, Overgaard K, Meden P, et al. Effects of citicolinecultures. Neurosci Lett 1999; 274 (2): 79-82combined with thrombolytic therapy in a rat embolic stroke

47. Liberatore G, Samson A, Schleuning W, et al. Unlike tissue-model. Stroke 1999; 30 (7): 1464-71type plasminogen activator, Desmodus rotundus (vampire bat)28. Carter LP, Guthkelch AN, Orozco J, et al. Influence of tissuesalivary plasminogen activator does not promote neurodegen-plasminogen activator and heparin on cerebral ischemia in aeration. Stroke 2003; 34: 537-43rabbit model. Stroke 1992; 23 (6): 883-8

48. Parsons MW, Barber PA, Chalk J, et al. Diffusion- and perfu-29. Del Zoppo GJ, Copeland BR, Waltz TA, et al. The beneficial

sion-weighted MRI response to thrombolysis in stroke. Anneffect of intracarotid urokinase on acute stroke in a baboon

Neurol 2002; 51 (1): 28-37model. Stroke 1986; 17 (4): 638-43

49. Schleuning WD. Vampire bat plasminogen activator DSPA-30. Lekieffre D, Benavides J, Scatton B, et al. Neuroprotection alpha-1 (desmoteplase): a thrombolytic drug optimized by

afforded by a combination of eliprodil and a thrombolytic natural selection. Haemostasis 2001; 31 (3): 118-22agent, rt-PA, in a rat thromboembolic stroke model. Brain Res

50. Diffusion-weighted imaging Evaluation For Understanding1997; 776 (1-2): 88-95Stroke Evolution (DEFUSE) Study [online]. Available from

31. Overgaard K, Sereghy T, Pedersen H, et al. Neuroprotection URL: http://www.stanford.edu/group/neurology/stroke/In-with NBQX and thrombolysis with rt-PA in rat embolic stroke. troduction.htm [Accessed 2003 Sep 29]Neurol Res 1993; 15 (5): 344-9

51. Rother J, Schellinger PD, Gass A, et al. Effect of intravenous32. Sakurama T, Kitamura R, Kaneko M. Tissue-type plasminogen thrombolysis on MRI parameters and functional outcome in

activator improves neurological functions in a rat model of acute stroke <6 hours. Stroke 2002; 33 (10): 2438-45thromboembolic stroke. Stroke 1994; 25: 451-6 52. Stroke trials directory: ECASS-III [online]. Available from

33. Sereghy T, Overgaard K, Boysen G. Neuroprotection by excita- URL: http://www.strokecenter.org/trials/TrialDetail.as-tory amino acid antagonist augments the benefit of throm- p?ref=475&browse=acute [Accessed 2003 Oct 28]bolysis in embolic stroke in rats. Stroke 1993; 24 (11): 1702-8 53. The Third International Stroke Trial (IST-3) [online]. Available

34. Zhang RL, Zhang ZG, Chopp M. Increased therapeutic efficacy from URL: http://www.dcn.ed.ac.uk/ist3/ [Accessed 2003 Sepwith rt-PA and anti-CD18 antibody treatment of stroke in the 29]rat. Neurology 1999; 52 (2): 273-9 54. Lijnen HR, Collen D. Staphylokinase. In: Backlin F, editor.

35. Rijken DC, Sakharov DV. Basic principles in thrombolysis: Fibrinolytics and antifibrinolytics. Berlin: Springer-Verlag,regulatory role of plasminogen. Thromb Res 2001; 103 Suppl. 2001: 425-491: S41-9 55. Sakharov DV, Lijnen HR, Rijken DC. Interactions between

36. Keyt BA, Paoni NF, Refino CJ, et al. A faster-acting and more staphylokinase, plasmin (ogen), and fibrin: staphylokinase dis-potent form of tissue plasminogen activator. Proc Natl Acad criminates between free plasminogen and plasminogen boundSci U S A 1994; 91 (9): 3670-4 to partially degraded fibrin. J Biol Chem 1996; 271 (44):

27912-837. Nagai N, Vanlinthout I, Collen D. Comparative effects of tissueplasminogen activator, streptokinase, and staphylokinase on 56. Vanderschueren S, Van Vlaenderen I, Collen D. Intravenouscerebral ischemic infarction and pulmonary clot lysis in ham- thrombolysis with recombinant staphylokinase versus tissue-ster models. Circulation 1999; 100 (25): 2541-6 type plasminogen activator in a rabbit embolic stroke model.

Stroke 1997; 28 (9): 1783-838. Booth NA. Regulation of fibrinolytic activity by localization ofinhibitors to fibrin (ogen). Fibrinolysis Proteolysis 2000; 14: 57. Thomas GR, Thibodeaux H, Errett CJ, et al. A long-half-life206-13 and fibrin-specific form of tissue plasminogen activator in


1010 Donnan et al.

rabbit models of embolic stroke and peripheral bleeding. territory stroke: determination of etiological, topographic, andStroke 1994; 25 (10): 2072-8 radiological outcome factors. Stroke 1998; 29 (12): 2529-40

58. Cannon CP, McCabe CH, Gibson CM, et al. TNK-tissue 76. Freitag HJ, Becker VU, Thie A, et al. Lys-plasminogen as anplasminogen activator in acute myocardial infarction: results adjunct to local intra-arterial fibrinolysis for carotid territoryof the Thrombolysis in Myocardial Infarction (TIMI) 10A stroke: laboratory and clinical findings. Neuroradiology 1996;dose-ranging trial. Circulation 1997; 95 (2): 351-6 38 (2): 181-5

59. Cannon CP, Gibson CM, McCabe CH, et al. TNK-tissue 77. Lewandowski CA, Frankel M, Tomsick TA, et al. Combinedplasminogen activator compared with front-loaded alteplase in intravenous and intra-arterial r-TPA versus intra-arterial ther-acute myocardial infarction: results of the TIMI 10B trial: apy of acute ischemic stroke: Emergency Management ofThrombolysis in Myocardial Infarction (TIMI) 10B investiga- Stroke (EMS) Bridging Trial. Stroke 1999; 30 (12): 2598-605tors. Circulation 1998; 98 (25): 2805-14 78. Ernst R, Pancioli A, Tomsick T, et al. Combined intravenous

60. Van de Werf F, Cannon CP, Luyten A, et al. Safety assessment and intra-arterial recombinant tissue plasminogen activator inof single-bolus administration of TNK tissue-plasminogen ac- acute ischemic stroke. Stroke 2000; 31 (11): 2552-7tivator in acute myocardial infarction: the ASSENT-1 trial: the 79. Zaidat OO, Suarez JI, Santillan C, et al. Response to intra-ASSENT-1 investigators. Am Heart J 1999; 137 (5): 786-91 arterial and combined intravenous and intra-arterial thrombo-

61. Torr SR, Nachowiak DA, Fujii S, et al. ‘Plasminogen steal’ and lytic therapy in patients with distal internal carotid arteryclot lysis. J Am Coll Cardiol 1992; 19 (5): 1085-90 occlusion. Stroke 2002; 33 (7): 1821-7

62. Chapman DF, Lyden P, Lapchak PA, et al. Comparison of TNK 80. Suarez JI, Zaidat OO, Sunshine JL, et al. Endovascular admin-with wild-type tissue plasminogen activator in a rabbit embolic istration after intravenous infusion of thrombolytic agents forstroke model. Stroke 2001; 32 (3): 748-52 the treatment of patients with acute ischemic strokes. Neuro-

surgery 2002; 50 (2): 251-963. Zhang RL, Zhang L, Jiang Q, et al. Postischemic intracarotidtreatment with TNK-tPA reduces infarct volume and improves 81. Hill MD, Barber PA, Demchuk AM, et al. Acute intravenous:neurological deficits in embolic stroke in the unanesthetized intra-arterial revascularization therapy for severe ischemicrat. Brain Res 2000; 878 (1-2): 64-71 stroke. Stroke 2002; 33 (1): 279-82

64. Kasai S, Arimura H, Nishida M, et al. Primary structure of 82. Keris V, Rudnicka S, Vorona V, et al. Combined intraarterial/single-chain pro-urokinase. J Biol Chem 1985; 260 (22): intravenous thrombolysis for acute ischemic stroke. Am J12382-9 Neuroradiol 2001; 22 (2): 352-8

65. Pannell R, Gurewich V. Pro-urokinase: a study of its stability in 83. The IMS Investigators. The Interventional Management ofplasma and of a mechanism for its selective fibrinolytic effect. Stroke (IMS) Study: safety results [abstract]. Stroke 2003; 34Blood 1986; 67 (5): 1215-23 (1): 247

66. Tebbe U, Windeler J, Boesl I, et al. Thrombolysis with recom- 84. Dirnagl U, Iadecola C, Moskowitz MA. Pathobiology of is-binant unglycosylated single-chain urokinase-type plasmi- chaemic stroke: an integrated view. Trends Neurosci 1999; 22nogen activator (saruplase) in acute myocardial infarction: (9): 391-7influence of heparin on early patency rate (LIMITS study): 85. Siesjo BK. Pathophysiology and treatment of focal cerebralliquemin in myocardial infarction during thrombolysis with ischemia. Pt II: mechanisms of damage and treatment. Jsaruplase. J Am Coll Cardiol 1995; 26 (2): 365-73 Neurosurg 1992; 77 (3): 337-54

67. Gurewich V, Liu JN. Intra-arterial pro-urokinase in ischemic 86. Globus MY, Busto R, Lin B, et al. Detection of free radicalstroke. Stroke 1998; 29 (6): 1255-6 activity during transient global ischemia and recirculation:

68. Reid H, Chan K, Thean P. Prolonged coagulation defect (defi- effects of intraischemic brain temperature modulation. Jbrination syndrome) in Malayan viper bite. Lancet 1963; I: Neurochem 1995; 65 (3): 1250-6621-6 87. Busch E, Kruger K, Allegrini PR, et al. Reperfusion after

69. Bell WR, Pitney WR, Goodwin JF. Therapeutic defibrination in thrombolytic therapy of embolic stroke in the rat: magneticthe treatment of thrombotic disease. Lancet 1968; I (7541): resonance and biochemical imaging. J Cereb Blood Flow490-3 Metab 1998; 18 (4): 407-18

70. Pollak VE, Glas-Greenwalt P, Olinger CP, et al. Ancrod causes 88. Yang Y, Li Q, Shuaib A. Enhanced neuroprotection and reducedrapid thrombolysis in patients with acute stroke. Am J Med Sci hemorrhagic incidence in focal cerebral ischemia of rat by low1990; 299 (5): 319-25 dose combination therapy of urokinase and topiramate. Neuro-

pharmacology 2000; 39 (5): 881-871. Prentice CR, Turpie AG, Hassanein AA, et al. Changes inplatelet behaviour during arvin therapy. Lancet 1969; I (7596): 89. Shuaib A, Yang Y, Li Q. Evaluating the efficacy of citicoline in645-7 embolic ischemic stroke in rats: neuroprotective effects when

used alone or in combination with urokinase. Exp Neurol72. Sherman DG, Atkinson RP, Chippendale T, et al. Intravenous2000; 161 (2): 733-9ancrod for treatment of acute ischemic stroke: the STAT study.

A randomized controlled trial: Stroke Treatment with Ancrod 90. Bowes MP, Rothlein R, Fagan SC, et al. Monoclonal antibodiesTrial. JAMA 2000; 283: 2395-403 preventing leukocyte activation reduce experimental neurolog-

ic injury and enhance efficacy of thrombolytic therapy. Neuro-73. Sherman DG. Ancrod. Curr Med Res Opin 2002; 18 Suppl. 2:logy 1995; 45 (4): 815-9S48-52

91. Zhang RL, Zhang ZG, Chopp M, et al. Thrombolysis with74. Trouillas P, Nighoghossian N, Getenet J, et al. Open trial oftissue plasminogen activator alters adhesion molecule expres-intravenous tissue plasminogen activator in acute carotid terri-sion in the ischemic rat brain. Stroke 1999; 30 (3): 624-9tory stroke: correlations of outcome with clinical and radiolog-

ical data. Stroke 1996; 27: 882-90 92. Zhang L, Zhang ZG, Zhang RL, et al. Postischemic (6-hour)75. Trouillas P, Nighoghossian N, Derex L, et al. Thrombolysis treatment with recombinant human tissue plasminogen activa-

with intravenous rtPA in a series of 100 cases of acute carotid tor and proteasome inhibitor PS-519 reduces infarction in a rat



model of embolic focal cerebral ischemia. Stroke 2001; 32 105. Minnema MC, Friederich PW, Levi M, et al. Enhancement of(12): 2926-31 rabbit jugular vein thrombolysis by neutralization of factor XI:

in vivo evidence for a role of factor XI as an anti-fibrinolytic93. Sanchez C, Alonso de Lecinana M, Diez-Tejedor E, et al.factor. J Clin Invest 1998; 101 (1): 10-4Treatment of embolic cerebral infarct via thrombolysis and

cytoprotection with U-74389-G in rats [in Spanish]. Rev 106. Klement P, Liao P, Bajzar L. A novel approach to arterialNeurol 1998; 27 (158): 653-8 thrombolysis. Blood 1999; 94 (8): 2735-43

94. Orozco J, Mendel RC, Hahn MR, et al. Influence of a ‘brain 107. Refino CJ, DeGuzman K, Schmitt D, et al. Consequences ofprotector’ drug 21-amino steroid on the effects of experimental inhibition of plasma carboxypeptidase B on in vivo throm-embolic stroke treated by thrombolysis. Neurol Res 1995; 17 bolysis, thrombosis and hemostatis. Fibrinolysis Proteolysis(6): 423-5 2000; 14: 305-14

95. Grotta J. Combination Therapy Stroke Trial: recombinant tis- 108. Nagashima M, Werner M, Wang M, et al. An inhibitor ofsue-type plasminogen activator with/without lubeluzole. Cer- activated thrombin-activatable fibrinolysis inhibitor poten-ebrovasc Dis 2001; 12 (3): 258-63 tiates tissue-type plasminogen activator-induced thrombolysis

96. Lyden P, Jacoby M, Schim J, et al. The Clomethiazole Acute in a rabbit jugular vein thrombolysis model. Thromb Res 2000;Stroke Study in tissue-type plasminogen activator-treated 98 (4): 333-42stroke (CLASS-T): final results. Neurology 2001; 57 (7):

109. Morita-Fujimura Y, Fujimura M, Gasche Y, et al. Over-1199-205expression of copper and zinc superoxide dismutase in trans-

97. Hacke W, Lees K, Krams M, et al. An adaptive dose response genic mice prevents the induction and activation of matrixfinding study of UK-279,276 in acute ischemic stroke [ab- metalloproteinases after cold injury-induced brain trauma. Jstract]. 28th International Stroke Conference; 2003 Feb 13-15; Cereb Blood Flow Metab 2000; 20 (1): 130-8Phoenix. Stroke 2003; 34 (1): 248

110. Gasche Y, Copin JC, Sugawara T, et al. Matrix metal-98. Lindgren A, Lindoff C, Norrving B, et al. Tissue plasminogen loproteinase inhibition prevents oxidative stress-associated

activator and plasminogen activator inhibitor-1 in stroke pa- blood-brain barrier disruption after transient focal cerebraltients. Stroke 1996; 27 (6): 1066-71 ischemia. J Cereb Blood Flow Metab 2001; 21 (12): 1393-400

99. Fisher M, Francis R. Altered coagulation in cerebral ischemia:111. Hartung HP, Kieseier BC. The role of matrix metalloproteinasesplatelet, thrombin, and plasmin activity. Arch Neurol 1990; 47

in autoimmune damage to the central and peripheral nervous(10): 1075-9system. J Neuroimmunol 2000; 107 (2): 140-7

100. Feinberg WM, Bruck DC, Jeter MA, et al. Fibrinolysis after112. Sumii T, Lo EH. Involvement of matrix metalloproteinase inacute ischemic stroke. Thromb Res 1991; 64 (1): 117-27

thrombolysis-associated hemorrhagic transformation after em-101. Weitz J, Crowther M. Direct thrombin inhibitors. Thromb Res bolic focal ischemia in rats. Stroke 2002; 33 (3): 831-6

2002; 106 (3): V275-84113. Lapchak PA, Araujo DM. Reducing bleeding complications102. Direct Thrombin Inhibitor Trialists’ Collaborative Group. Di-

after thrombolytic therapy for stroke: clinical potential ofrect thrombin inhibitors in acute coronary syndromes: princi-metalloproteinase inhibitors and spin trap agents. CNS Drugspal results of a meta-analysis based on individual patients’2001; 15 (11): 819-29data. Lancet 2002; 359 (9303): 294-302

103. Mikulski A, Elg M, Gustafsson D. The effects of oral andintravenous direct thrombin inhibitors on the size of Correspondence and offprints: Dr Geoffrey A. Donnan, Na-photochemically induced cortical infarction in rats. Thromb

tional Stroke Research Institute, Austin & RepatriationRes 2001; 101 (6): 477-82Medical Centre, 300 Waterdale Rd, West Heidelberg, Victo-104. Sakharov DV, Plow EF, Rijken DC. On the mechanism of theria 3081, Australia.antifibrinolytic activity of plasma carboxypeptidase B. J Biol

Chem 1997; 272 (22): 14477-82 E-mail: [email protected]


Documents

Can the Time Window for Administration of Thrombolytics in Stroke be Increased?