Bladin et al. appear to be advocating a return to eminence-based medicine, highlighting the many bodies that have approved tissue plasminogen activator (tPA) stroke thrombolysis.[1] There are plenty of examples of how these same groups have got it wrong, such as the VioxxTM debacle (rofecoxib, Merck & Co., Inc., Whitehouse Station, NJ, USA) or steroids for spinal cord injury. Harmful treatments have been promoted by professional bodies and consensus guidelines.[2] The correspondents misinterpret the Australasian College for Emergency Medicine position statement on i.v. thrombolysis for ischaemic stroke which explicitly states: ‘There is insufficient evidence for stroke thrombolysis to be considered a standard of care.’[3] This is in keeping with other international emergency medicine colleges. Indeed, the International Journal of Stroke has a leading article by a stroke physician, Tony Furlan from the United States, who is not convinced by IST 3 either and highlights its flaws.[4] The ‘group-think’ of Bladin et al. fails to comprehend that critical evaluation is essential for medicine.

The authors cite the Ingall et al. paper as supportive of their opinion, but a careful reading of this paper highlights anything but that view.[5] Those authors were cautious because absence of proof is not proof of absence. Given the small number of patients overall, there is essentially no power whatsoever to estimate statistically the impact of such confounding (from the imbalance in baseline stroke severity) in multiple tiny subgroups. Thus, the Ingall et al. reanalysis merely found no proof of confounding, a foregone conclusion, rather than strong evidence of its absence.[5]

Meta-analysis is promoted as the pinnacle of evidence. However, recent reports highlight that we cannot necessarily rely on meta-analysis.[6] Increasingly, there are reports that meta-analyses are missing over half of all patient data. When the meta-analysis is redone with inclusion of the missing data, the result can be completely different, for example, reboxetine and oseltamivir.[7, 8] Researchers have had their faith in published reports shaken, and of course, methodologic flaws contaminate the literature and are not eliminated by meta-analysis.[7] As a result, Godlee and Loder of the BMJ have called for urgent action to restore the integrity of the medical evidence base.[8] Other respected researchers highlight the current epidemic of false claims and that claimed research findings are likely to represent accurate measures of the prevailing bias.[9, 10] More recently, the integrity of the ECASS III result (one of the only two purportedly positive stroke lysis studies) has been challenged, which further discredits the evidence base for thrombolysis in stroke, upon which meta-analysis relies.[11]

While preparing this response, I was forwarded an email from the daughter of a man who received tPA as advocated by Bladin et al. Tragically, several hours later, he suffered a catastrophic intracerebral haemorrhage and died. The daughter wrote: ‘Whatever happened to the motto doctors used to live by “cause no harm”?’ Primum non nocere.

It is a sobering fact that thrombolysis increases the likelihood of an early death. Whether it results in meaningful improvements in neurological function of the survivors and whether we can target a group that will benefit is unknown. I whole-heartedly agree that further research, especially replication, is essential. However, we still need to prove that thrombolysis or mechanical thrombectomy provides a benefit to any group of patients.[12] This is the inconvenient truth.

Competing interests

None declared.


  1. Top of page
  2. References