more_vert
IT was reported last month that the Australian Competition and Consumer Commission (ACCC) had instituted claims against a NSW homeopath for posting allegedly misleading material on her website.
Most readers would agree that people making false or exaggerated claims about any therapeutic product should be held to account.
How well, then, do we apply those principles to our own practice? While we all support the principles of evidence-based practice (where appropriate evidence exists), are there areas where our own profession may be making claims beyond the evidence?
This made me look at the evidence, claims and evolving practice in the use of thrombolytic therapy for stroke.
While there is no comparison between the scientific principles of revascularisation and the anti-science of homeopathy, there is a common underlying principle. We need to be realistic with our claims and honest with our patients — and each other — about what the evidence actually shows.
Many of us remember the initial studies that established the place of thrombolysis in heart attack. The findings were reproduced over and over, and the evidence accumulated exponentially.
The criteria to thrombolyse were relatively straightforward, many people benefitted, and very few were significantly harmed.
It was natural that clinicians treating stroke — also a disease of vascular occlusion — would look for similar results to help their patients. While we know that stroke patients benefit in the long-term from admission to a specialised stroke unit, it makes sense to look for a more active and immediate therapy.
As interventional cardiology began to take over, thrombolytic therapy was looking for a new role.
So, what is the stroke thrombolysis research? Have the initial positive results been replicated in a huge database of strong evidence of benefit over harm?
Despite the trend for stroke thrombolysis to be seen by neurologists as a standard of care, the answer is, unfortunately, no.
A news article in MJA InSight last year illustrates the dilemma. The story, headed “Consider thrombolysis in all stroke patients”, reported on two papers in The Lancet suggesting that a wider range of stroke patients could potentially benefit from thrombolysis. It is in the readers’ comments on the article where critical commentary of the literature occurs.
The trial that first put stroke thrombolysis on the world stage was that by the National Institute of Neurological Disorders and Stroke, published in 1995. This work was not replicated in a major study until 2008, when the European Cooperative Acute Stroke Study (ECASS 3) was published.
ECASS 3 was a manufacturer-supported randomised controlled trial (RCT), with just over 800 patients enrolled. Despite randomisation, there were 5% more cases with severe stroke in the placebo group — which could account for the entire benefit found. Major symptomatic haemorrhages occurred in 2.4% of the thrombolysed group vs 0.2% of the placebo group.
To this point, nine of the 11 major acute stroke trials had demonstrated either no benefit or harm.
And what is the nature of the benefit? A study published in JAMA showed a small improvement in function at 30 days in patients administered thrombolysis between 3 and 5 hours after symptom onset (greatest in those with the mildest strokes), and no improvement in mortality (more in the thrombolysed group died in the early phase).
Therefore, to meet truth in advertising standards in 2013 thrombolysis in stroke requires this explanation: “There may be some long-term benefit from clot-busters in acute stroke but it’s not great. We don’t know which patients will benefit most, and who will die from it. It’s no miracle.”
Dr Sue Ieraci is a specialist emergency physician with 30 years’ experience in the public hospital system. Her particular interests include policy development and health system design, and she has held roles in medical regulation and management.
Posted 18 March 2013
Click here to read a news report on why experts are demanding a national strategy to improve stroke care, including wider use of thrombolysis
You sum the situation up well, Francis. If the evidence base were strong, it would sell itself – without the need for badgering. On the other hand, we know that Stroke Units work well, and don’t kill anyone.
The issue really hits home when our ED is being asked to “get with the times” and accept a “proven” therapy, being chastised for appalling thrombolysis rates, with the implication that we somehow don’t care, or suffer from some form of ignorance. These acusations have been levelled at a room full of concerned doctors wanting more evidence before prescribing a therapy with a proven harm profile, by an increasingly irate stroke physician. During the rise of thrombolysis for myocardial infarction, we saw the Cardiologists own the problem, and its research, and eventually hand over the administering of the treatment (and its consent process) to Emergency Physicians – in what was eventually a patient-centred collaboration. This was possible because the evidence base was consistently showing the same signal. Stroke physicians/neurologists are approaching the stroke thrombolysis issue differently, often berating Emergency Physicians for being obstructive to research translation. The difference being that the research is not clear with regards to this treatment for this condition (except on the issue of early deaths…) The problem is that immense volumes of funding that should be going to the OT’s, Physios, nurses and Speechies in a new Stroke Unit, are being earmarked for expensive ambulance diversion and telemedicine programs, and 24 hour Stroke Physician on-call payments. Given the vested interests involved in the generation of the “evidence base” for rTPA in stroke, I provide only a half-hearted apology for my cynicism relating to this new paradigm…
I wonder why “Reading the literature” is unwilling to identify themselves, as everyone else is willing to do? The problem with his/her comment is that thrombolysis for stroke is indeed being pushed as a standard of care around Australia. Some neurologists (and emergency physicians, I might add) are aggressively pushing for major system changes such as ambulance diversion to stroke centres, with early thrombolysis a major reason for such early diversion instead of timely transfer after assessment and stabilisation in the nearest ED. This has major cost and safety implications, all for a treatment that has a seriously flawed evidence-base. The Stroke Foundation seems to consider stroke thrombolysis to be safe (and therefore a standard of care) in some settings. What many of us are arguing is that without further randomised controlled trials this is wrong and is putting our patients in harm’s way.
If “Reading the Literature” would care to “read the article”, he/she would find that the quote “Most thinking doctors should be alarmed by the proposition that thromblysis be considered a standard of care without further research. ” is correctly attributed to another of the commenters above, not to myself. Perhaps “R the L” might also like to retract “Apparently to disagree with her is to invite a response that neurologists can’t read the literature. The fact is they wrote it.” Yes, some neurologists wrote up the studies, but I don’t know how many neurologists critically reviewed the results. Like everyone, I want the best for patients. I would love there to be an effective and safe acute treatment for strokes. Unfortunately, we don’t have it yet – even in places with organised stroke care or stroke units.
With reference to Richard Gerraty’s comments “For doctors in such a can-do specialty as emergency medicine the opposition to a proven treatment like tPA for stroke is extremely puzzling.” I have heard and seen written down similar statements elsewhere and I have a real problems with the words “proven treatment”. To me “proven” is the language of spin doctors not medical doctors or scientists. I was taught that to prove something scientifically is impossible. Proofs exist only in mathematics and logic, not in science. Mathematics and logic are both closed, self-contained systems of propositions, whereas science is empirical and deals with nature as it exists. The primary criterion and standard of evaluation of scientific theory is evidence, not proof. Proofs are not the currency of science, hence we practice evidence based medicine and not “proof” based medicine.
So if we get back to the basics, the scientific method requires the development of a null hypothesis (thrombolysis does not improve the functional outcome in stroke) and if exhaustive experimental evidence can not support the null hypothesis then the hypothesis is accepted – but not proven. It is impractical to strictly use the scientific method in its purest form in human clinical trails. This I accept. But if the essence of the scientific method is then transposed into to the area of medical trials, where there are no null hypotheses, then every clinician giving a treatment (especially one that can harm some patients) is ethically bound to make every attempt to find holes in the evidence supporting a hypothesis and try to disprove it. Only by doing this can the medical profession be confidant we are doing the best for our patients and accept a treatment based on the evidence. Emergency Physicians can and do apply these basic scientific principles to our daily practice. This is why we are the can-do specialty.
Finally, I do not understand why those who challenge the evidence for thrombolysis in stroke are seen are opponents? I question the strength of the evidence and I strongly object to any treatment in medicine being labeled “proven”. But I am in no way ideologically opposed to thrombolysis and would gladly embrace this treatment if the evidence base was strong enough. Surely we should be not be talking in black and white terms like “proven” or “unproven”, “supporters” or “opponents” and start using the language of science – and talk about the evidence, accepting that evidence comes in shades of grey.
For someone who urges doctors to read the literature carefully, the claim made by Sue Ieraci that
“Most thinking doctors should be alarmed by the proposition that thrombolysis be considered a standard of care …”
is curious. In fact thrombolysis is not considered a standard of care, and the Stroke Foundation does not support it’s use outside of appropriately staffed and equipped settings. Thrombolysis for stroke is not included in its stroke care bundle precisely because it is not a standard of care and cannot be considered safe in many clinical settings. Perhaps Sue Ieraci should read the Stroke Care Bundle at http://www.nhmrc.gov.au/_files_nhmrc/publications/attachments/cp116_comp… which states “Thrombolysis has not been included in the care bundle because it is currently not recommended for routine use in hospitals without dedicated and organised stroke care or stroke units. Inclusion of thrombolysis as a component of the bundle would not be in line with the care bundle approach, which is to develop a resource that can be implemented in all situations.”
Apparently to disagree with her is to invite a response that neurologists can’t read the literature. The fact is they wrote it.
When you look at the neuroradiology literature you can see some are moving to IA tPA. When you look at studies with either angiograms/dopplers looking at return of cerebral A’s flow after IV thrombolytics, the more distal the lesion it appears there is an increased rate of recanalisation (52% good outcome in MCA M2 vessels v 18% ICA vessels). My take on the whole debate is that we are lumping all strokes together (small vessels/ big ICA occlusion) and hence the difficulty in getting any impressive results compared to data from MI thrombolysis.
Stroke is a terrible disabling disease that is so frustrating to manage. I understand the enthusiasm for an active intervention. Just what that is I am not so sure. In my regional hospital we may see 1-2 /mo who may fit thrombolytic criteria and yet we see hundreds of CVAs a year, where an effective and well resourced stoke unit would have a far better impact.
Refs
Early recanalization after intravenous administration of recombinant tissue plasminogen activator as assessed by pre- and post-thrombolytic angiography in acute ischemic stroke patients.Stroke 2007
Site of Arterial Occlusion Identified by Transcranial
Doppler Predicts the Response to Intravenous
Thrombolysis for Stroke. Stroke 2007
IST-3, the largest ever randomised trial of t-PA in stroke (n=3035), was a largely unblinded trial in which functional outcome at 6 months was self-assessed by patients by either postal or telephone questionnaire in most instances. Both of these factors would tend to bias in favour of t-PA. Despite this there was no significant difference in mortality at 6 months and no difference in the predefined outcome (proportion of patients with OHS 0-2). There was a significant increase in deaths at 7 days in the t-PA group (OR 1.6 95% CI 1.22-2.08). A secondary ordinal analysis showed some improvement in functional outcome at 6 months, but caution is justified when a secondary outcome is used to claim efficacy, particularly in the face of prevailing biases. It is not reasonable to say this is simply about offering patients choice since human nature tends to overestimate the benefits and underestimate the risks. In addition prognosticating final neurological outcome early after stroke onset is notoriously difficult (except in the most severe cases in which t-PA is contraindicated because it unequivocally results in poorer outcome). The biases of the treating doctor have a huge influence on patient decisions. Under the ‘first do no harm’ principle it is difficult to justify a clear risk of early death on the basis of a possible increase of 1 point on the OHS/mRS later on. This is in sharp contrast to MI where thrombolytics unequivocally reduce short and long-term mortality despite a small risk of harm. Stroke doctors see the devastating consequences of this disease over the long term. Emergency physicians have a unique view of the competing claims upon finite healthcare resources. Both groups want to do the best for their patients. Productive collaborations are required to urgently research which patients will benefit based upon stroke aetiology, neuro-imaging and interventions targeting the culprit vessel. In the meantime the considerable resources invested in thrombolytics for stroke may be better directed towards proven therapies such as increasing availability of multi-discipinary stroke unit care.
Ian Wilson – IST-3 – published last year, gave thrombolysis between 0 and 6 hours post-onset, and found no discernible relationship between timing and drug effect. The results looked good in the first three hours, but then harmful for the next 90 minutes, and then good again for the next 90. (It is worth looking closely at the data). IN addition, therapy did not reduce death or dependence at 6 months post-treatment. You say ” I think it is reassuring that we are not saving people from death to end up highly dependent.” Not only are we “not saving people from death”, we are causing their death, without reducing the number that are highly dependent. We don’t actually know whether the ones that are killed by the therapy are the ones who would have been the most severely affected, following long-term rehab. All we know is that the modest benefit that has been shown in some studies was shown in the mildest strokes. I would recommend that people read the body of each of the papers – not just the abstracts, and see what methodology was used and what the data actually showed.
I think a distinction needs to be made between trials that allowed thrombolysis out to 6hours vs those that used a 3hr (or 4.5hour) time frame. There seems to be a clear correlation between early thrombolysis and better outcomes. Crudely stating that x number of trials were negative vs y number positive lumps these together and is not informative. Thrombolysis is not recommended beyond 4.5hours where I have worked and is only used between 3-4.5hours with caution. NINDs (and the <3hour subgroup of IST-3 and the Cochrane meta-analysis) showed for that for patients thrombolysed within 3 hours more will have a favourable outcome and that despite an increase in early death this is made up for by 6 months at which point there is no difference. I would take that treatment if I had a stroke. I think it is reassuring that we are not saving people from death to end up highly dependent. ECASS III seemed to quite convincingly extend the time frame to 4.5 hours. The benefit beyond 4.5hours should be seen as a completely separate issue to be debated (..the data for which, so far looks rather unconvincing).
Richard Gerraty (“For doctors in such a can-do specialty as emergency medicine the opposition to a proven treatment like tPA for stroke is extremely puzzling.”) In reply, I say that for doctors with such a high level of training as neurologists, I find the willingness to embrace such a dodgy evidence base as being “proven” to be perplexing if not downright disturbing. Sue and others have provided very simple and convincing critiques of the evidence base. These easily accessed critiques (google “stroke thrombolysis lifeinthefastlane”) should be carefully considered by all neurologists. Most thinking doctors should be alarmed by the proposition that thromblysis be considered a standard of care without further research. Many emergency physicians stand ready to collaborate and get to the bottom of this issue. We are indeed a “can do” specialty.
Richard Gerraty – I was waiting for someone to mention “a can-do specialty as emergency medicine”. Have you ever wondered why emergency physicians were champions of cardiac thrombolysis, but not stroke thrombolysis? It’s the evidence, combined with no vested interests. You say the Wardlaw meta-analysis “was still positive for death and disability.” But Wardlaw said this “Thrombolytic therapy increased the risk of symptomatic intracranial haemorrhage (OR 3.49, 95% CI 2.81 to 4.33) and death by three to six months after stroke (OR 1.31, 95% CI 1.14 to 1.50). Treatment within three hours of stroke appeared more effective in reducing death or dependency (OR 0.71, 95% CI 0.52 to 0.96) with no statistically significant adverse effect on death (OR 1.13, 95% CI 0.86 to 1.48).” IST-3 (Lancet 2012) found an excess of 11% deaths within 7 days in the rt-PA group. By 6 months, similar numbers, in total, had died (408 [27%] in the rt-PA group vs 407 [27%] in the control group)(the control group dies of other causes. So, no benefit in deaths, an excess of deaths caused directly by the treatment, in exchange of a modest benefit in outcome (best in the mildest stroke). Clearly some people do benefit, but not the most severe strokes, and we don’t know which people will benefit and which will die – even under trial protocols. If can-do emergency physicians are cautious about this evidence, shouldn’t everyone be? What would we have to gain by just being oppositional?
“Anonymous” says that “There exists (sic) systematic (and therefore less prone to bias) ways to assess clinical trial data..” Yes – that’s called meta-analysis – which is what I discussed in the article. The limitations of article length prevented me from a long dissertation, but the comments on this and related articles contain a lot more of the evidence. The comments of “Anonymous” and Elizabeth Farrier appear to misunderstand the nature of the benefits found in those studies that did find benefit. The choice is not about risking death for a dramatic improvement in function – it is about risking death when a mild stroke might recover completely vs almost completely. If you are giving your patients the choice of death vs a dramatic improvement for a severe stroke, you are not practising truth in advertising – thrombolysis doesn’t give those sorts of results.
The 2012 Cochrane Review (Wardlaw et al) methodology required the inclusion of all thrombolytics including streptokinase and yet it was still positive for death and disability. The recent trials of tPA at 0.9mg/kg and without the addition of aspirin show that thrombolytic therapy does not ‘kill’ patients early. Haemorrhage is an ever-present risk of this therapy, but the risk is acceptable on the basis of the trials of the current dose and drug, and on the basis of the large European registry, SITS-MOST. Medicine is quite used to weighing up front risk versus long term benefit. Opponents of this therapy claim that there should be as many patients in stroke trials as were included in coronary thrombolysis trials, ignoring the fact that tPA is much more effective for stroke than for heart attack, and less patients were needed to prove benefit. More than 7000 patients contributed to the 2012 meta-analysis, and for AMI the thousands and thousands of patients randomized after the first 5000 were actually in trials of an already proven therapy.
For doctors in such a can-do specialty as emergency medicine the opposition to a proven treatment like tPA for stroke is extremely puzzling. Such opponents are shrinking in number. Collaboration between neurologists and emergency physicians, in the countries where that specialty exists, is now the norm for early stroke therapy.
Sue’s article, doesn’t represent a sound way of getting to the truth in this matter, but rather, reads as a pre-formed opinion with selective use of evidence and persuasive language techniques. Her position may or may not be correct, but how could you tell.
There exists systematic (and therefore less prone to bias) ways to assess clinical trial data and as medical doctors, we really are obliged to use those techniques when we arrive at a considered view. Isn’t the lack of scientific rigour the basis for an ACCC pursuit of homeopathists? That Sue uses arguments such as “…It is in the readers’ comments on the (research) article where critical commentary of the literature occurs…” may be strong literary argument, but it is weak in science – the best that readers can usefully add is an opinion, ie, the lowest form of scientific evidence.
The argument that might well be made regarding thrombolysis and ischaemic stroke is that different people put different weights on the benefits and risks of this intervention, and may therefore arrive at a different decision with the same information presented. So, shouldn’t thrombolysis be offered to the appropriate people and the patients/surrogate be allowed to make up their own mind (ie, informed consent)?
I think we should also consider that this is a decision to be made with the patient and their family. Sometimes the risk of death may outweight the prospect of months of rehabilitation followed by a life without work or the ability to participate in pre stroke activities. Thrombolysis in many cases works and patients are discharged home from hospital to continue on with their life as they knew it.
Advocates of thrombolysis in stroke often point to the results of the meta-analyses showing an overall slight improvement in long term functional outcome despite increases in early haemorrhage and death. Of the twelve trials to date, four were stopped early due to harm, thus diluting the effect of these trials in the pooled analyses. Of the two “positive” trials, the NINDS study also had baseline differences in stroke severity which accounted for the differences in outcomes and ECASS-3 used a novel and clinically questionable primary outcome measure. While there may be an undefined subset of patients in whom very early treatment (<3h) provides net benefit, a dispassionate assessment of the evidence does not support the widespread use of this therapy at the present time. Thanks Sue for raising this important public health issue. We are long overdue for a rational debate based upon science, not spin.
(Thanks, Dr Relf). “Confused” – the Wardlaw et al meta-analysis, published in Lancet 2012, combined the results of 12 trials, totalling just over 7000 patients. More people died in the thromboylsis arm. “Numbers of deaths within 7 days were increased (250/2807 [8·9%] vs 174/2728 [6·4%], 1·44, 1·18—1·76; p=0·0003), but by final follow-up the excess was no longer significant (679/3548 [19·1%] vs 640/3464 [18·5%], 1·06, 0·94—1·20; p=0·33). Symptomatic intracranial haemorrhage (272/3548 [7·7%] vs 63/3463 [1·8%], 3·72, 2·98—4·64; p<0·0001) accounted for most of the early excess deaths.” IN terms of outcome, the difference between a modified Rankin Scale of 0-2 vs 0-1 is from “No significant disability despite symptoms; able to carry out all usual duties and activities” (1) to “Slight disability; unable to carry out all previous activities, but able to look after own affairs without assistance.”. IN other words, the therapy killed people initially, but, in the long term, people with mild strokes had some (small) benefit in function. This is not of the nature of totally dependent people suddenly becoming fully recovered, as if often portrayed. OF course, some people with quite dense paresis do recover in front of your eyes – even without thrombolysis. Those are TIAs.
Excellent scientific appraisal. Drs do EBM sometimes; and refreshing when its reported and discussed. I like Sue’s no nonsense style. Keep it up.
I’m slightly confused. Also in this same edition of MJA Insight, there is a stroke article which claims “only 7% of ischaemic stroke patients had received thrombolysis treatment, despite evidence that such treatment resulted in an extra 10 independent survivors for every 100 patients who received it”, referencing the Lancet. Yet here (and my previously held belief) the claim is the opposite.
Any comment??
I think truth in advertising, should mandate the statement:
“There may be some long-term benefit from clot-busters in acute stroke but we aren’t sure, the benefit doesn’t seem to be large, and we have no way of knowing if you will be the one who benefits or dies as a result of treatment. However, what we DO know beyond any doubt is that clot-busters increase the chance of you dying earlier than you otherwise would have.”
2.4% haemorrhage risk is often downplayed as a small risk but cerebral haemorrhage from thrombolysis is invariably CATASTROPHIC and untreatable.
Spot on. What distinguishes medical pratitioners (specialist or otherwise) from homeopaths is that we should be albe to give our patients expert advice to support them in making an informed treatment decision. Risks AND benefits neeeds to be discussed.