To the Editor: The assertion of Davis and Batmanian and their colleagues that thrombolytic treatment for ischaemic stroke is reasonably safe and highly effective1,2 is not supported by the primary randomised trials of its use.3

Is tissue plasminogen activator (tPA) safe? It remains undisputed that none of the trials for thrombolysis in stroke have shown any mortality benefit.4 All of the trials have shown increases in symptomatic intracranial haemorrhage; in the NINDS trial, the increase was from 0.6% in the placebo arms to 6.4% in the treatment arms — a 1000% relative increase — and 45% of those with symptomatic bleeds died.5 Presumably, any mortality benefit from opening blocked arteries is lost because of the increased mortality from intracranial bleeding. Therefore, the drug is not safe.

Is tPA effective? Thrombolysis for acute myocardial infarction was assessed in tens of thousands of patients in many independent studies, with virtually all showing clear mortality benefit. By comparison, the stroke thrombolysis literature is full of negative studies, with only one positive result. The NINDS trial4 stands alone as the only randomised controlled trial (RCT) providing positive evidence for thrombolysis for stroke. Ignoring criticisms of its interpretations and methodology, of which there are many, it had fewer than 600 patients and its results have not been reproduced independently.

Breaches of protocol continue to be published. Batmanian and colleagues gave patients tPA after 180 minutes despite all the evidence saying this has no benefit, justifying it by saying the decision had been made at 170 minutes.2

Davis and colleagues1 based their claims of safety and efficacy on registries, subgroup analyses, meta-analyses, and expert panels all based on the same single RCT — the NINDS trial.4 This is low-grade evidence for a potentially lethal therapy.

It is time a major RCT was done to repeat the NINDS trial and finally determine whether its result was a statistical anomaly or a real effect. There is no shortage of stroke patients — Batmanian et al found that 14% of patients were eligible for this therapy.2

The paucity of evidence for thrombolysis for stroke does not justify rushing patients to stroke centres, bypassing perfectly good hospitals in the hope of finding some of the 3% of patients eligible for treatment, of whom one in eight (0.38% of all stroke patients) would theoretically have a better neurological symptom score if given thrombolysis.6,7 I feel that it is a waste of time and effort, and a danger to patients, to focus all resources on supplying a potentially lethal therapy that is often incorrectly used and provides a marginal benefit.