Three Part Question
In [patients diagnosed with acute ischaemic stroke] does [thrombolysis] improve [mortality and functional outcome]?
A 49 y/o women is brought to the emergency department by ambulance. She was having lunch with friends when she suddenly collapsed at the table. On arrival, she is aphasic and has a marked weakness of her right side. Her partner attends and informs you of her medical history which consists of treated hypertension and nil else of note. You arrange for an urgent CT scan and contact the on-call stroke physician. The patient has attended within 1h of onset of symptoms and you wonder if she would be a candidate for thrombolysis. You mention this possibility to the patient's partner and he bombards you with a series of specific and highly appropriate questions regarding the risks and benefits of thrombolysis in patients with stroke. You point out that the diagnosis needs to be confirmed by CT scan before this treatment can even be considered and then cunningly deflect his questions towards the stroke physician who has just arrived in the department.
You are aware that thrombolysis for acute ischaemic stroke is considered beneficial for a certain group of patients within strict inclusion and exclusion criteria, but feel that you should know more about the outcomes of a treatment that is usually administered in your department.
EMBASE and MEDLINE 1996-09/09 using the OVID interface.
The Cochrane Database of Systematic Review.
The search was made for articles containing original data from randomised controlled trials comparing thrombolytic agents against placebo.
[((brain or cerbr$ or cerebell$ or vertebrobasil$ or hemispher$ or intracran$ or intracerebral or infratentorial or supratentorial or mca$ or anterior circulation).ti. AND (ischaemia or ischemia or infarc$ or thrombo$ or emboli$ or occlus$).ti.) OR stroke.ti. OR cva.ti. OR cerebrovascular accident.ti.] AND [thromboly$.ti. OR fibrinoly$.ti. OR (tpa or t-pa or rt-pa or rtpa or plasminogen).ti. OR (anistreplase or streptodornase or streptokinase or urokinase or pro?urokinase or rpro?uk or lumbrokinase or duteplase or lanoteplase or pamiteplase or reteplase or saruplase or staphylokinase or streptase or alteplase).ti.] LIMIT to humans, English language and randomized controlled trial.
The Cochrane Database was searched using the terms 'thrombolysis' and 'stroke'.
1318 papers were found from the literature search, 3 papers were considered relevant to the three-part-question.
1 Cochrane review was considered relevant to the three-part-question.
|Author, date and country
||Study type (level of evidence)
|Wardlaw et al.|
|Systematic review and meta-analysis of 18 clinical trials with a total of 5727 randomised patients. Only 5675 patients are included due to missing data. Not all of the trials provided data for all of the outcomes that were looked at by the systematic review.||Systematic review and meta-analysis.||Death from all causes in first 7-10 days.||14.9% in treatment groups vs. 9.4% in placebo groups. OR 1.81 (95% CI 1.46-2.24)||1 trial (NINDS 1995) provided the bulk of available data for the outcomes of patients who received treatment within 3h of onset of symptoms. |
|Fatal intracranial haemorrhage||5.2% in treatment groups vs. 0.9% in placebo groups. OR 4.34 (95% CI 3.14 – 5.99)|
|Symptomatic haemorrhage||8.7% in treatment groups vs. 2.5% in placebo groups. OR 3.37 (95% CI 2.68-4.22)|
|Death from all causes up to end of scheduled follow up period||18.2% in treatment group vs. 15.2% in placebo groups. OR 1.33 (95% CI 1.15-1.53) |
|Death or dependency ||53.3% in treatment groups vs. 58% in placebo groups OR 0.84 (95% CI 0.75-0.95) |
|Death or dependency at end of scheduled follow in patients treated within 3h||49.7% in placebo groups vs. 60.3% in control groups. OR 0.66 (95% CI 0.53-0.83)|
|Death in patients treated within 3h at end of scheduled follow up period||20.9% in treatment groups vs. 19.4% in placebo groups OR 1.13 (95% CI 0.86-1.48)|
|Ogawa et al.|
|114 patients aged 20-75y, with focal neurological signs corresponding to the middle cerebral artery distribution confirmed by angiography as an acute ischaemic stroke. Various exclusion criteria included any clotting problems and an NIHSS >22. Patients with a history of a previous CVA were also excluded.||RCT
Patients were randomised to receive intra-arterial urokinase or standard treatment.||(Primary outcome) modified Rankin scale of 0-2 at 90 days. ||49.1% in the treatment group vs 38.6% in the control group. OR 1.54 (95% CI 0.73-3.23); p=0.345.||The study was ended prematurely due to the licensing of intravenous rtPA for stroke thrombolysis in Japan. The treatment was not blinded though assessment of clinical outcome was.|
|Mortality rates ||5.3% in the treatment group vs. 3.5% in the control group; p=0.647|
|Intracranial haemorrhage ||9% in the treatment group vs. 2% in the control group; p=0.206. |
|Davis et al.|
|101 patients with acute ischaemic stroke presenting for treatment 3-6h following onset of symptoms. Patients recruited from 15 centres in the UK, Belgium, Australia and New Zealand. Patients had to have a NIHSS greater than 4 and a premorbid modified Rankin score of 2 or less.||RCT
Patients were randomised following a CT scan to receive placebo or 0.9 mg/kg (up to a maximum of 90mg) of alteplase intravenously. 52 patients were randomised to alteplase and 49 to placebo.||Mortality||25% in the treatment group vs. 14.2% in the placebo group. OR 2.0 (95% CI 0.72-5.53)||Small study. Significant numbers of patients lost to follow up, 6 in treatment group, 1 in placebo group. If the worst outcome is assumed for these patients then this significantly worsens mortality data.|
|Symptomatic ICH||7.7% in the treatment group vs. 0% in the placebo group.|
|Good outcome (modified Rankin score ≤ 1)||34.6% in the treatment group vs. 24.5% in the placebo group. OR 1.6 (95% CI 0.69-3.9)|
|Good outcome (modified Rankin Score ≤2)||42.3% in the treatment group vs. 40.8% in placebo group. OR 1.06 (95% CI 0.48-2.35)|
|Hacke et al.|
|821 adult patients with acute ischaemic stroke presenting for treatment within 3-4.5h of onset of symptoms. Various exclusion criteria including NIHSS >25, haemorrhage on CT scan or suspected haemorrhage despite a normal CT scan and hypertension.||RCT
Patients were randomised to receive 0.9mg/kg (up to max of 90mg) of intravenous alteplase or placebo. Patients were examined by an assessor, who was blind to the treatment, at 1h, 2h and 24h after administration and then 7, 30 and 90 days.||Favourable outcome (modified Rankin score of 0-1)||52.4% in the treatment group vs. 45.2% in the placebo group. OR (1.34 95% CI 1.02 to 176); p=0.04||Study protocol was modified after the first 228 patients had been enrolled to extend the inclusion period from 4 to 4.5h. This was due to a suggestion from pooled analysis of other trials suggesting a benefit up to 4.5h and slow recruitment.
The main concern is that the incidence of previous stroke in the placebo group was 14.1% while in the treatment group it was 7.7%. The modified Rankin score for patients prior to this stroke is not available.|
|Favourable outcome (modified Rankin score of 0-2)||66.5% in the treatment group vs. 61.6% in the placebo group. OR 1.30 (95% CI 0.95-1.78); p=0.11|
|Incidence of intracranial haemorrahge||27% in the treatment group vs. 17.6% in the placebo group; p=0.001|
|Incidence of symptomatic ICH (defined as a increase in NHISS score of 4 or greater)||2.4% (10 patients) in the treatment group vs. 0.2% (1 patient) in the treatment group. OR 9.85 (95% CI 1.26-77.32); p=0.008|
|Mortality rates at 90 days.||6.7% in the treatment group vs. 7.7% in the placebo group. OR 0.86 (95% CI 0.51-1.46)|
It is difficult to draw conclusions from the available evidence regarding thrombolysis in stroke. While it appears to have a beneficial outcome in terms of avoidance of the combination of dependency and death this is largely down to one trial. The NINDS trial (1) published in 1995 looked at a total of 624 patients with ischaemic stroke who were treated within 3h. The patients were randomised to receive 0.9mg/kg (up to a maximum of 90mg) alteplase or placebo. This demonstrated that patients receiving alteplase were more likely to have a modified Rankin score of 0 or 1 at 3 months than those who received placebo. It also demonstrated that despite an increase in symptomatic and fatal intracerebral haemorrhage the overall mortality was lower for the patients who received treatment.
The ECASS trial (2), conducted in Europe at around the same time as the NINDS trial, included 620 patients with ischaemic stroke who received treatment up to 6h following onset of symptoms and had rather different results. The patients were randomised to receive 1.1mg/kg (up to a maximum dose of 100mg) of alteplase or placebo but in this trial the mortality rates were significantly higher in the treatment group (22% vs. 15.6%) with no significant improvement in functional outcome. There was no significant difference in the patients who received treatment before or after 3h although the majority of patients were treated after this time period.
Further unpublished data was made available to the FDA in the application for the licence to use alteplase for treatment of ischaemic stroke and this is available from the FDA website. http://www.fda.gov/downloads/Drugs/DevelopmentApprovalProcess/HowDrugsareDevelopedandApproved/ApprovalApplications/TherapeuticBiologicApplications/ucm080832.pdf. The data available from the NINDS study shows a trend towards imbalance in the baseline severity of the stroke with a larger number of patients in the treatment group having a milder stroke. This fact prompts the authors of the Cochrane review to comment that the results of this study should be regarded with ‘EXTREME CAUTION’ with their deliberate use of capital letters.
The authors of the Cochrane report conclude that the data may justify the use of thrombolytic therapy using rTPA in experienced centres in highly selected patients but that the data do not support the widespread use of thrombolytic therapy in routine clinical practice at this time.
The FDA felt differently and granted a licence for alteplase to be used in patients presenting with acute stroke within 3h who have had haemorrhage excluded by CT or some other form of imaging modality with similar sensitivity.
In the UK NICE has also approved the use of alteplase for acute ischaemic stroke for patients who present early enough to receive the treatment within 3h (http://www.nice.org.uk/TA122). This was with the caveat that it should be used by a physician trained and experienced in the management of acute stroke and only in centres with facilities that enable it to be used in full accordance with its marketing authorisation. The data from the first ECASS trial was excluded from consideration due to the fact that dose higher than the proposed therapeutic dose was used (1.1mg/kg vs. 0.9mg/kg).
In the trials published since 2003 there are similar results to those found by the Cochrane Review. There is a tendency to move away from the middle ground with patients more likely to have a good outcome or die when treated with alteplase than the patients who receive placebo. There are a few issues with the published data that is available.
The papers tend to report a modified Rankin score of 0 or 1 as a good outcome. It may be argued that while the difference between 0 (no disability) and 6 (death) is obvious, the difference between 1 (no significant disability despite symptoms) and 2 (slight disability but able to look after all affairs without assistance) is more subjective. The patients’ pre-morbid modified Rankin scores were not provided which may be relevant given a fair proportion of them have had strokes prior to the presenting event. The largest study, ECASS III, had an incidence of previous stroke of 14.1% in the placebo group while only 7.7% in the treatment group.
It is not clear from reading these papers what the cause of death was in the patients who received placebo. While the treatment groups invariably had a higher rate of fatal intracerebral haemorrhage the overall mortality rates were similar. If these patients have died as a result of an extension of their original ischaemic stroke, bearing in mind that patients with more severe strokes have been excluded from enrolment in the more recent studies, then it is easier to balance the risk of causing more fatal haemorrhagic stroke with that of reducing fatal ischaemic stroke. If, however, these patients died from unrelated causes then one may feel less confident in recommending thrombolysis.
A further issue, one that commonly arises in medical research, is that trials tend to be carried out at specialist centres. Patients tend to have better outcomes when involved in studies, and we know that patients with stroke treated in specialised stroke units have better outcomes so results from research in these specialised centres may not be applicable to general clinical practice.
Another issue is the heterogeneity of the patients in the studies. This is less apparent in the inclusion criteria or demographics of included patients than the outcomes, with quite different mortality rates even in the placebo groups. Compared with trials on thrombolysis for myocardial infarction the total number of patients involved in these trials is very small. This makes it very difficult to draw conclusions about the overall risks of benefit and harm and apply them to an individual patient. For example, even using the data from the meta-analysis the overall mortality rate for patients receiving placebo was 15.2% while in patients who received placebo within 3h the mortality rate was 19.4%.
Editorial comment from the EMJ Mar 2010; The Cochrane review by Wardlaw et al has since been updated, although the general conclusions remain the same. The need for further research is acknowledged by the fact that there is an ongoing Third International Stroke Trial (IST-3), which may answer some of the outstanding questions (http://www.dcen.ed.ac.uk/ist3/).
Clinical Bottom Line
This area is crying out for further research. The over-riding impression is that some patients receive benefit from thrombolysis while other patients are harmed, and indeed may die as a result of this treatment. The question that needs to be answered with robust, good quality evidence is how we decide which patients are which.
In keeping with NICE guidance, local treatment guidelines regarding thrombolysis need to be agreed with the local stroke physicians.
Level of Evidence
Level 1 - Recent well-done systematic review was considered or a study of high quality is available.
- Wardlaw JM, del Zoppo GJ, Yamaguchi T et al. Thrombolysis for acute ischaemic stroke Cochrane Database of Systematic Reviews 2003, Issue 3. Art. No.: CD000213
- Ogawa A, Mori E, Minematsu K et al. Randomized Trial of Intraarterial Infusion of Urokinase Within 6 Hours of Middle Cerebral Artery Stroke The Middle Cerebral Artery Embolism Local Fibrinolytic Intervention Trial (MELT) Japan Stroke 2007;38;2633-2639.
- Davis SM, Donnan GA , Parsons MW et al. Effects of alteplase beyond 3 h after stroke in the Echoplanar Imaging Thrombolytic Evaluation Trial (EPITHET): a placebo-controlled randomised trial Lancet Neurology 2008; 7: 299–309.
- Hacke W, Kaste M, Bluhmki E. Thrombolysis with Alteplase 3 to 4.5 Hours after Acute Ischemic Stroke N Eng J Med 2008;359:1317-29 .
- The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group Tissue Plasminogen Activator for Acute Ischemic Stroke N Eng J Med 1995;333:1581-7
- 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