New drugs, or new trials of current drugs, for the treatment of acute ischaemic stroke? added 10/01/01

Lancet Sept 1, 2001

The treatment of acute ischaemic stroke continues to be a challenge. For every step forward there seems to be many steps sideways. The report by Philip Bath and colleagues in this issue of The Lancet, of a lack of effectiveness of the low-molecular-weight heparin, tinzaparin, compared with aspirin, adds to the plethora of recent "negative" trials of anticoagulants, neuroprotective agents, and other treatments for ischaemic stroke. This result begs the question: "Where are new drug treatments for stroke going to come from?".

Promising new strategies include:
- early arterial recanalisation by intra-arterial pro-urokinase,
- intra­venous defibrinogenating agents, such as ancrod,
- intravenous glycoprotein IIb receptor antagonists, and oral platelet ADP-receptor antagonists;
- minimisation of cytotoxic brain damage by intravenous magnesium,
- neutrophil-inhibitory factor, and anti-inflammatory drugs; and
- prevention of recurrent stroke by agents, such as statins that may stabilise atherosclerotic plaque.

However, before throwing the baby out with the bath water (no pun intended), it is important to examine why so many trials of current drugs have been "negative".

Have we misunderstood the cause and pathogenesis of ischaemic stroke, and thus developed inappropriate drugs? Or misunderstood the pharmacology of existing drugs and how they interact with each other? Or developed invalid animal models of ischaemic stroke for testing drugs? Or failed to recognise modest but important treatment effects because random error was not minimised (for example, by prematurely terminating clinical trials based on futility analyses)? Furthermore, why does the treatment of an occluded cerebral artery and its sequelae seem to be so much more difficult and less successful than the treatment of an occluded coronary artery?

The treatment of acute myocardial infarction (MI) is now generally standard throughout the world,¹ but the treatment of acute ischaemic stroke remains inconsistent. Some patients, but by no means all, are given aspirin,² fewer are admitted to stroke units³ or given heparin,⁴ and, in countries in which it is approved, intravenous thrombolysis⁵ is given to fewer than 2% of patients.⁶

There are several possible explanations for this inconsistency:

·         The cause of  ischaemic stroke is heterogeneous. Whereas almost all cases of acute MI are caused by in situ atherothrombosis (ie, plaque erosion or fissure followed by platelet aggregation, and fibrin thrombus), this process accounts for only about one-fifth of cerebral infarction.⁷ Another two-fifths are caused by atherothromboembolism, one-fifth by emboli from the heart (eg, fresh thrombus, organised thrombus, calcium, bacteria, fragments of prosthetic valves or tumour), and one-fifth by other causes (eg, arterial dissection, arteritis, procoagulant states).⁷ It is unrealistic to expect effective treatments for MI to be as widely applicable and effective for all causes of ischaemic stroke. The overall effect of any treatment is likely to be modest, at best. The difficulty is compounded by the lack of reliable and valid criteria for classifying ischaemic stroke into subtypes based on cause.

• The response of the brain to ischaemia differs from that of the myocardium. Although ischaemia causes a time and flow dependent cascade of complex chemical events⁸ in both the brain and the heart, glutamate seems to be a major mediator of cell death in the brain, whereas it seems to enhance myocardial protection (it is used in cardioplegic solutions to boost ATP production via the Krebs citric-acid cycle). Furthermore, neurons are entirely dependent on glucose metabolism for energy and have less potential for anaerobic metabolism via the glycolytic pathway than do myocardial cells, which also use fatty acids as a fuel. Consequently, the time available to rescue ischaemic, functionally impaired, but surviving brain tissue (the ischaemic penumbra) is probably much shorter than that for ischaemic myocardium. Although this interval may be up to 17 h,⁹ as it is for MI,¹ it is probably only a few hours in most cases. Consequently, ischaemic stroke is even more of a medical emergency than MI, and treatment probably needs to be given very early—"time is brain". How prompt treatment can be effected in practice is another challenge.
  
• The response of the brain to haemorrhagic transformation of the infarct is different, and less forgiving, than that of the myocardium. Although effective treatments for MI¹ are also effective for ischaemic stroke, their favourable effects in ischaemic stroke are compromised substantially, and in the case of heparin completely, by their adverse effects on the brain. Treating 1000 patients with an ischaemic stroke with thrombolysis within 6 h of onset saves 44 (95% CI 15-73) overall from death or dependency but the benefit would be much greater if it were not at the cost of an excess of 70 (58-83) symptomatic intracranial haemorrhages and 44 (34-54) fatal intracranial haemorrhages.⁵ Similarly, treating 1000 ischaemic stroke patients with aspirin within 48 h of onset prevents nine recurrent ischaemic strokes but causes two haemorrhagic strokes and four extracranial haemorrhages.² And, treating 1000 patients with ischaemic stroke with systemic anticoagulants (heparins, heparinoids) prevents nine recurrent ischaemic strokes but causes nine haemorrhagic strokes, with no net improvement in death or dependency.⁴
  
• The potential effectiveness of some treatments for stroke (compared with MI) are not widely appreciated. Although the treatment of acute ischaemic stroke seems to be less successful than for acute MI, it may not be. Thrombolysis within 6 h of acute MI prevents 30 deaths per 1000 patients treated¹ whereas tissue-plasminogen activator given intravenously within 3 h of onset of ischaemic stroke prevents 140 from death or dependency.⁵ Furthermore, early (within 48 h) use of aspirin for ischaemic stroke prevents 12 per 1000 treated from death and dependency,² and care in an organised stroke unit prevents a further 66 from death and dependency.³ These effective treatments, particularly the latter two—which are safe—need to be delivered to many more appropriate patients than are currently receiving them.⁶
  
• Important treatments for ischaemic stroke are likely to have modest effects that can be reliably identified or excluded only by means of very large, blinded randomised controlled trials involving tens of thousands of  patients, rather than the tens, or hundreds, of patients enrolled in the many seemingly "negative" stroke trials. More than 58 600 patients with MI were evaluated in clinical trials of thrombolysis before modest but important treatment effects were reliably identified,  and  the  treatment has become standard  practice.¹ Similarly, 41 325 patients with acute ischaemic stroke were randomised in clinical trials of aspirin to reliably identify a mild but worthwhile and cost-effective treatment effect (treating 1000 patients prevents 12 from death or dependency).² However, only 5216 patients with ischaemic stroke due to several causes have been studied in clinical trials of thrombolysis.⁵ Although a systematic review⁵ of these trials has concluded that thrombolysis is effective, it remains uncertain in whom it is effective, in whom it is ineffective, and in whom it is dangerous. The best thrombolytic agent, dose, half-life, and route of administration, and the most effective concomitant neuroprotective, antithrombotic, and antihypertensive regimen are also unknown. Only one small trial has evaluated the effect of one (rather large) dose of a thrombolytic agent combined with aspirin,⁵ and one very small trial has studied the effects of thrombolysis combined with a neuroprotective agent in a total of 89 patients.¹⁰ There have been no trials of the combination of thrombolysis, antiplatelet therapy, and one or more neuroproteqtive drugs which act on different parts of the  cascade  of ischaemic  neuronal  injury. Neuroprotective drugs need access to the ischaemic tissue before they can be  expected to work,  and recanalised arteries need to remain recanalised and not re-thrombose. Thrombolytic agents dissolve red-blood-cell thrombus, but predispose to rethrombosis through platelet activation, release of bound thrombin, and activation of factors V and VIII by plasmin. Although streptokinase combined with aspirin was associated with an excess risk of intracranial bleeding in one small underpowered trial,' the use of one or more antiplatelet agents with different mechanisms of action combined with a much lower dose of a thrombolytic agent (eg, one  quarter  of current  doses)  may be  safer  and surprisingly effective.

There is as much promise with the existing drugs, perhaps used in different doses and combinations, as with new drugs of the future. There is good evidence that thrombolysis and aspirin are both more effective than hazardous, and that heparin is as effective as it is dangerous. An immediate challenge is to reliably identify' which stroke patients are likely to benefit from which dose and from which of these effective therapies, and which patients are likely to be harmed, so that risky but effective treatments can be confidently targeted to only those who are likely to benefit. However, this identification can be achieved only by proper evaluation in very large randomised trials, such as the International Stroke Trial-3 (IST-3).'¹ In IST-3, the plan is that, within 6 h of stroke onset, 6000 patients with ischaemic stroke will be randomly assigned intravenous tissue-plasminogen activator or placebo. Patients will be randomised only if their clinician is uncertain about the relative risks and benefits of thrombolysis in that patient. Because clinicians have different thresholds of uncertainty, there will be a wide variation in the types of patients enrolled, and consequently, there will be important information about the consistency of treatment effect in a broad range of patients. If successful, the individual patient data from this trial, when added to the current database of 5216 patients from 17 previous trials, should have the statistical power to reliably identify which clinical and neuroimaging features best predict patients who are likely to benefit, not benefit, or be harmed by thrombolysis. In the meantime, other promising treatments must also be evaluated appropriately, and not discarded prematurely on the basis of futility and financial analyses. The results of all trials should be updated regularly in the Cochrane database of systematic reviews, and when conclusive, translated into the professional and public domain, and implemented in practice.

Graeme J Hankey

Department of Neurology, Royal Perth Hospital, Perth WA 6001, Australia (e-mail: gjhankey@cyllene.uwa.edu.au )

1 Collins R, Peto R, Baigem C, Sleight P. Aspirin, heparin, and fibrinolytic therapy in suspected acute myocardial infarction. N EnglJMed 1997; 336: 847-60.

2 Gubitz G, Sandercock P, Counsel! C. Antiplatelet therapy for acute ischaemic stroke. In: The Cochrane Library, Issue 2. Oxford: Update Software, 2001.

3 Stroke Unit Trialists' Collaboration. Organised inpatient (stroke unit) care for stroke. In: The Cochrane Library, Issue 2. Oxford: Update Software, 2001.

4 Gubitz G, Counsel! C, Sandercock P, Signorini P. Anticoagulants for acute ischaemic stroke. In: The Cochrane Library, Issue 2. Oxford: Update Software, 2001.

5 Wardlaw JM, del Zoppo G, Yamaguchi T. Thrombolysis for acute ischaemic stroke. In: The Cochrane Library, Issue 2. Oxford: Update Software, 2001.

6 Katzan IL, Furlan AJ, Lloyd LE, et al. Use of tissue-type plasminogen activator for acute ischemic stroke. The Cleveland area experience. JAMA 2000; 283s 1151-58.

7 Warlow CP, Dennis MS, van Gijn J, et al. Stroke: a practical guide to management. 2nd ed. Oxford: Blackwell, 2000.

8 Lee JM, Zipfel GJ, Choi PW. The changing landscape of ischaemic brain injury mechanisms. Nature 1999; 399 (6738 suppl): A7-A14,

9 Baron JC. Mapping the ischaemic penumbra with PET: implications for acute stroke treatment. Cerebrovasc Dis 1999; 9: 193-201.

10 Grotta JC. Combination therapy stroke trial: rt-PA +/- lubeluzole. Stroke 2000; 31: 278 (abstr).

11 Hand P, Lindley R, Wardlaw J, Sandercock P. The third International Stroke Trial (IST-3). Cerebrovasc Dis 2001; 11 (suppl 4): 35 (abstr).