Pro debate: Alteplase for Stroke Thrombolysis
Dr Tom Daubeny
Firstly the background...
Epidemiology of Stroke:
Cause of 8.5% of UK Deaths, leading cause of adult disability
Broadly people who’ve had a stroke will have:
- 26% No Disability
- 25% Mild - Moderate Disability
- 27% Severe Disability
- 22% Death
The Outcome Measures used to analyse treatments for Stroke:
Modified Rankin Score (mRS):
0 – No symptoms, 1 – No significant disability, 2 – Slight disability, 3 – Moderate disability (can walk unassisted), 4 - Moderately severe disability, 5 – Severe disability, 6 – Dead
10 Activity measures scored to give a total of 100:
Feeding, Bathing, Grooming, Dressing, Bowel Control, Bladder Control, Toilet use, Transfers, Mobility, Stair use
Glasgow Outcome Scale
1 – Death, 2- Persistent vegetative state, 3- Severe disability (permanent need for help with daily living), 4-Moderate disablility, 5- Low disability
National Institutes of Health Stroke Scale (NIHSS)
11 outcome measures to give a total score of 42
Level of consciousness, Horizontal eye movement, Visual field deficit, Facial palsy, Motor deficit upper limb, Motor deficit lower limb, Ataxia, Sensory deficit, Dysphasia, Dysarthria, Sensory inattention
And the Current UK guidelines for Thrombolysis of Stroke
- NICE Recommendations
- “Alteplase for the treatment of acute ischaemic stroke 2007” (reviewed 2010)
- “Stroke – Diagnosis and initial management of acute stroke and TIA 2008”
- Given by physician trained in neurological care and in a well organised stroke service (staff trained in thrombolysis, immediate access to imaging and re-imaging)
- Given within 3hours of symptoms, post imaging to exclude intracerebral haemorrhage (ICH)
- Age range 18-80
- 0.9mg/kg (up to 90mg) over 1 hour with 10% initial bolus
So what is the evidence supporting these guidelines of giving Alteplase within 3 hours of onset of stroke symptoms?
1/Evidence to show significantly improved outcome if given within 3 hours
2/NO mortality difference (despite initial increased ICH rate with treatment)
3/ Large post marketing and phase 4 studies support these findings of improved outcome and no mortality difference
National Institute of Neurological Studies (NINDS) Randomised Control Trial 1995
- 2 part double blind trial of 624 patients
- Outcome at 24hours (291patients) and outcome at 3months (333patients) analysed
- Favourable outcome at 3months for all outcome measures (Barthel index, modified Rankin Scale, Glasgow outcome scale and NIHSS) with odds ratio 1.7 (1.2-3.6)
- 30% more likely to have minimal or no disability at 3 months
- Symptomatic ICH 6.4% vs. 0.6%, no difference in mortality
- NNT 3 to change improve 1 in mRS
- Data reanalysed in 2004 – odds ratio 2.1
European Cooperative Acute Stroke Study 3 (ECASS 3) 2008
- Multicentre double blind RCT in 19 European Countries
- 821 patients looking at 3-4.5 hours
- Follow up at 3 months
- Better outcome up to 4.5 hours – Modified Rankin Score 0-1 at 90 days Odds Ratio 1.34 (1.02 – 1.76)
- NNT 15
International Stroke Trial 3 (IST 3) 2012
- Multi centred randomised open trial 3035 patients from 12 countries
- Allocated 0.9mg/kg alteplase or standard of care within 6 hours
- Follow up at 18months
- No difference mortality
- Significantly increases survival free of handicap – odds ratio 1.28 (1.03-1.57)
Lancet 2012 Systematic review and meta-analysis
- 12 RCTs alteplase within 6 hours included- 7012 patients
- Final follow up between 1-6 months
- Increase favourable outcome 5.5% (3.3-7.7) on modified Rankin scale
- Increase in survival free of dependency 4.2% (1.9-6.6%)
- If given within 3hours: 9% (4.6-13.5) increase in survival free of dependency and 8.7% (4.6-12.8) increase in favourable outcome
- Symptomatic ICH – 5.8% (4.9-6.8) absolute increase, no significant mortality at final follow up
- Review commented no significant heterogeneity between trials for any of these outcomes
Safe Implementation of Thrombolysis – Stroke Monitoring Study (SITS-MOST) 2002-2006
- Prospective, open, multicentre, multinational, observational monitoring study established as a condition by the European Union for licensing
- 6483 patients 285 centres 14 European countries
- 39% with no or mild disability at 3 months vs. 29% controls
Get With the Guidelines (GWTG) 2012
- USA National Stroke Registry 2003-2012
- 58,353 patients from 1395 hospitals alteplase within 4.5 hours
- Faster Onset to treatment time associated with reduced in hospital mortality, reduced symptomatic ICH, increased independent ambulation at discharge and increased discharge home
Safe Implementation of Treaments in Stroke - International Stroke Thrombolysis Register (SITS-ISTR)
- International monitoring registry for auditing the safety and efficacy of thrombolysis across 19 countries
- 29,618 patients, 25,279 within 3 hours
- Confirms safety profile up to 3 hours and supports safety up to 4.5 hours post symptom onset
Canadian Alteplase for Stroke Effectiveness Study 2005
- 60 centres across Canada
- 2247 patients, 1135 within 3 hours
- Concludes stroke thrombolysis is a safe and effective therapy
What is the future for Alteplase?
Use of scoring to help individualise risk or select subgroups of patients most likely to benefit e.g. DRAGON Score, HAT score and iScore
Lower risk of intracerebral haemorrhage patients with posterior circulation stroke
J Neuro Sci 2013• 518 patients treated with Alteplase: 434 anterior circulation – 8.1% sICH vs 84 posterior circulation – 1.2% sICH
4 factors identified increasing ICH risk in NINDS analysis
Baseline NIHSS > 20; Age > 70; Ischemic changes present on initial CT; Glucose > 16.7 mmol/L
0 risk factor 1.8%; 1 risk factor 4.9%; >2 risk factors 21.2%
Increasing the benefit of alteplase
Concurrent use of agents that will work on older clots
Agents to protect against haemorrhage e.g. trials looking at minocycline
Directing Alteplase to site of occlusion
Alteplase has good evidence for beneficial outcome in functional status post ischaemic stroke if given within 3 hours of symptom onset.
It has been given a Level A recommendation for use by the major international institutions examining its efficacy: NICE (National Institute of Clinical Excellence), SIGN (Scottish Intecollegiate Guidelines Network), ESO (European Stroke Organisation), ACEM (Australasian college emergency medicine), AAEM (American Academy of Emergency Medicine)/ AAN (American Academy of Neurology/ American Stroke Association (ASA), Candadia Stroke Network.
There is no mortality difference with placebo in long term follow up. This is despite an early increased intracerebral haemorrhage rate and indicates that with targeted therapy a mortality benefit may also be possible.
Stroke Thrombolysis The ‘No’ Campaign
I got the chance to campaign against stroke thrombolysis. I first went through the trials that have shown a benefit for thrombolysis.
The first of these was NINDS. The first part of this trial looked at improvement at 24 hours following tPA compared with placebo, when given within 3 hours of stroke onset. When previously thrombolysing myocardial infarctions we remembered that we would see a quick improvement in ST changes, and therefore if tPA is going to make a difference in strokes, we would expect to find an improvement within 24 hours. Unfortunately there was no significant difference.
It is common teaching that a TIA lasts an average of 10 minutes and rarely greater than an hour. It was therefore of note that in the sub-group analysis of patients that had symptoms for 91-180 minutes prior to thrombolysis, or placebo, 39% and 37% respectively had resolution of their symptoms or an improvement in ≥4 points on the NIHSS score at 24 hours. Therefore the next time a stroke doctor claims benefit from tPA within 24 hours of treatment, point out that the evidence does not support that claim and that ~40% of patients will make an improvement at this stage with, or without thrombolysis.
NINDS II investigated the proportion of patients with minimal or no deficit at 3 months. They showed an absolute risk reduction of 13% following tPA compared to placebo. What they didn’t initially tell you was that there were big differences in the baseline characteristics. In the 91-180 minute subgroup, the tPA group had an extra 15% of patients with a NIHSS score of 0-5, whereas the placebo group had an extra 9.2% of patients with a NIHSS score of >20. With these huge differences in baseline characteristics it’s surprising that they only found an ARR of 13% with tPA.
The only other RCT to have shown benefit from thrombolysis is ECASS III. The time window for thrombolysis in this study was 3-4.5 hours. They found that the proportion of patients with a modified rankin score (mRS) of 0 or 1 at 90 days was 52.4% in the thrombolysis group vs. 45.2% in the placebo group. (P=0.04, OR 1.34 (1.02-1.76)). However, the placebo group had a 14.1% incidence of previous stroke vs. 7.7% in the placebo group. Some of these patients are unlikely to have had a mRS of 0/1 at the onset of the trial, which seems a little unfair! By using a cut-off of 0/1 for the mRS, this meant that death (mRS of 6) was counted the same as slight disability (mRS 2 = unable to carry out all previous activities, but able to look after own affairs without assistance) If the outcome had been the proportion of patients with a mRS of 0-2 at 90 days the results would have no longer been significant. Therefore thrombolysis was not significantly more likely than placebo to leave you independent following a stroke. And considering that 27% of patients in the tPA group had an intracerebral bleed, this study wouldn’t convince me that thrombolysis is the answer!
We then reviewed the trials that have either shown harm or no benefit from thrombolysis.
There have been 3 streptokinase trials. Two of these were stopped early due to excess mortality (MAST Europe and ASK). Whereas MAST Italy showed no change in the primary outcome of combined 6 month fatality and severe disability; but an excess 10 day mortality. Of note a recent Cochrane review found no difference in affect when comparing one thrombolytic agent with another. Although there was not a direct comparison of streptokinase vs. tPA.
Looking at the tPA/alteplase trials:
ATLANTIS A, included patients within 6 hours of stroke onset. They found an improvement in the NIHSS score in the tPA group but unfortunately had to be stopped early due to excess mortality in the tPA group (23% vs. 7%, P=0.01).
They then tried again with ATLANTIS B, which was predominately within 5 hours of stroke onset. This time they just showed a trend to excess mortality in the tPA group (10.9% vs. 6.9%, P=0.08)
ECASS I included patients up to 6 hours. They found no difference in their primary outcome of a mRS score but again an excess mortality in the tPA group.
ECASS II included patients predominately within 3-6 hours of stroke onset. They found no benefit in the primary outcome of mRS at 90 days, but at least the mortality was not significantly increased!
IST-3 was supposed to be the trial that finally put this matter to rest with the inclusion of over 3000 patients, within 6 hours of stroke onset. They found no difference between the tPA and the placebo group for the primary outcome of being alive and independent at 6 months. This was despite the trial being hugely biased in favour of the tPA group. The majority of the trial was non-blinded. We know that non-blinded trials are more likely to find a positive outcome for an intervention when compared when blinded trials. Of interest, the trial was initially blinded and for those patients there was a trend to improvement in the placebo group, where as in the non-blinded group there was a trend to improvement in the tPA group. The non-blinding meant that patients were treated differently with patients in the tPA group being much more likely than those in the placebo group to have been admitted to HDU. Therefore they are likely to have had their low oxygen saturations and hypotension picked up at an earlier stage, thereby preventing secondary brain injury. After being given this ‘wonder drug’ and having the care and attention of the HDU staff the patients may have been more likely to engage in their rehabilitation and they may have given more favourable answers to the questionnaire that was sent out to them (Which was how the primary outcome was determined for a number of patients) Therefore when a non-blinded trial of over 3000 patients, that is biased in favour of the intervention finds no improvement in the primary outcome, I am concerned that a higher quality trial may have found harm related to the intervention
In summary there have been 2 trials that have shown a benefit in the primary outcome following tPA. Both of these have methodological flaws. Whereas 3 trials have had to be stopped because of an excess mortality in the thrombolysis group. Another study showed an excess mortality but wasn’t stopped early. And 5 trials have shown no benefit in the primary outcome. Therefore it is clear that stroke thrombolysis has not been proven to be of benefit and currently