To the Editor: I respect the opinion of Hoffman1 and others who have recently expressed concerns about the use of tissue plasminogen activator (tPA) in patients with ischaemic stroke. However, it is critical that these views do not dampen the enthusiasm for better stroke treatments. The Therapeutic Goods Administration recently approved tPA for ischaemic stroke patients with symptoms of less than 3 hours’ duration. This approval has catalysed collaboration between stroke units around Australia and may provide benefits to patients who do not receive tPA, by improving clinical pathways, as well as a more coordinated approach to treatment.

Internationally acclaimed Australian stroke experts have already rebutted the arguments of opponents of the National Institute of Neurological Disorders and Stroke (NINDS) trial, 2,3 and an independent reanalysis of the NINDS data showed tPA to be more effective than originally reported.4 It is time to stop bickering about the NINDS trial5 and move forward, by focusing on a collaborative effort between emergency departments and stroke teams. In the near future, there will surely be better data on new-generation thrombolytics and other agents; all will operate on similar “time is brain” protocols, as did the NINDS trial.

Obviously, more data would be useful, but when clinicians are faced with patients with acute stroke today they must give the patients and their families the facts about tPA, regardless of their personal opinion. These are best expressed in terms of the absolute risk reduction for disability seen in the NINDs trial. It is interesting to note the widespread discomfort with the 6.4% risk of intracerebral haemorrhage (half of which were fatal) in this trial of a “medical” therapy. Surgeons quote similar morbidity and mortality risks every day to patients undergoing procedures such as coronary artery bypass grafting. Additionally, many surgical procedures became established on much less solid evidence than the NINDs trial, and yet there is no outcry. For example, many thousands of carotid endarterectomies were performed before the publication of controlled-trial data.

In properly selected patients in expert hands, we now have a treatment that works. Its risks should not be underestimated, but should also be placed into context when compared with other powerful therapies for serious illnesses. Patients presenting with acute ischaemic stroke in Australia today have the chance to benefit from tPA, but future patients will benefit even more, partly because of the process that is being undertaken to institute pathways for using this drug.

In reply: I agree with Blacker that enthusiasm generated by thrombolytic therapy could lead to benefit for stroke patients, independent of whether the therapy is actually useful, or even whether they receive it. However, I believe there are better ways to encourage rational stroke care.

I also acknowledge that some experts believe thrombolysis is proven to be beneficial if used in accordance with NINDS (National Institute of Neurological Disorders and Stroke) guidelines. Others believe the question is not yet settled and are concerned that widespread adoption of the guidelines will lead to more harm than good. It is one thing to enthuse about possible benefits of a treatment, particularly for a devastating disease for which we have traditionally had little to offer. It is entirely different to embrace this therapy, even though it may be harmful overall, simply out of this frustration.

I disagree with Blacker that the best way to inform patients about thrombolysis is “in terms of the absolute risk reduction for disability seen in the NINDS trial”. Regardless of concerns about the trial itself, it is always foolhardy to accept data from one small study as “the truth”. Many other trials have found far worse results — it would be equally inappropriate to base all estimates on their worst outcomes. Finally, there is strong evidence that, even if NINDS-like “success” could be achieved under optimal circumstances, routine use in the community might well lead to overall harm.

I am also cautious about Blacker’s comparison of thrombolysis and some surgical procedures in terms of adverse effect rates. No single adverse effect rate is or is not acceptable for all interventions — a 90% rate of intracerebral haemorrhage could be acceptable in a disease with 100% mortality, while a 2% rate would be completely unacceptable in a disease with no long-term morbidity. Furthermore, the adoption of many surgical interventions in the absence of persuasive evidence does not justify repeating this mistake.

Of course we must routinely make decisions without definitive evidence, based on our best estimates of benefits and harms. We should never take such decisions lightly, and should in general embrace the precautionary principle, which tells us not to adopt new therapies without reasonable evidence of their safety (especially when any benefit is likely to be extremely small).

Although some advocates support thrombolysis based on current knowledge, my reading of available evidence is far less sanguine. That is why I continue to argue that this treatment should not be introduced into routine practice until far better evidence of its benefit outweighing its harm becomes available. Given the possibility that this treatment will harm stroke patients overall, and the likelihood that any potential benefit is very limited for the stroke population as a whole, I ask once again, why is so much being made of so little?