To the Editor: The recent report on the use of a stroke thrombolysis protocol by Batmanian and colleagues1 raises many issues. The most recent guidelines regarding thrombolytic therapy for ischaemic stroke recommend that “If thrombolytic therapy is to be used . . . consultation with a neurologist or stroke physician is essential before instigating therapy. Strict adherence to the inclusion and exclusion criteria is important . . .”2 Hence, I find the integral role of a neurologist in Batmanian and colleagues’ protocol reassuring.

However, consent to therapy in this setting is not simple. In the protocol, they informed the patient and/or next of kin about the risks and benefits of thrombolysis, but did they seek informed consent? The benefits of thrombolytic therapy are controversial, and equipoise exists.

Reflecting this, a multicentre trial3 is underway that includes patients who meet the eligibility criteria of Batmanian et al’s protocol. The evidence cited by Batmanian and colleagues in support of thrombolysis was based on small randomised trials published more than a decade ago.4 Further, the summary results of the meta-analysis they cited are associated with significant heterogeneity and lose statistical significance with the inclusion of additional results from observational studies5 (Box).

Moreover, the more recently published audits and the analysis that led to the number-needed-to-treat estimate cited by Batmanian et al were based on comparisons with a reference group dominated by the placebo group of the NINDS (National Institute of Neurological Disorders and Stroke) trial.6 This trial has been criticised for likely faulty randomisation resulting in baseline disparities unfavourable to the placebo recipients. These projections of benefit based on comparisons with a small and non-representative placebo group from a trial published over a decade ago are misleading. A calculation of numbers needed to treat based on such limited evidence, particularly for a soft endpoint such as assessment of disability, is hazardous and potentially misleading.7

There are other important issues associated with this therapy that need to be resolved. For example: Are mortality rates increased in groups receiving tissue plasminogen activator therapy in non-teaching hospitals?;8 Is this therapy safe in older people?; and finally, does this therapy not have the proven mortality risk associated with streptokinase in randomised trials in patients with stroke?4

Thrombolysis for stroke requires more evidence from randomised trials, not more protocols. Trials including older people are especially needed, given the age-dependent mortality risk associated with this therapy.

Odds ratios for fatal outcome from published cohorts of patients receiving tissue plasminogen activator therapy within 3 hours of onset of acute ischaemic stroke

The addition of two recent observational cohorts, the German Stroke Register and the Cleveland experience, to the meta-analysis of six controlled trials as reported in Wardlaw et al4 increases the summary odds ratio from 0.97 (95% CI, 0.69–1.36) to 1.21 (95% CI, 0.93–1.57). (Figure reproduced from Hurley5 with permission from Wiley-Blackwell Publishing Ltd.) ATLANTIS = Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke. ECASS = European–Australasian Cooperative Acute Stroke Study. NINDS = National Institute of Neurological Disorders and Stroke.