The IST-3 trial

It is the biggest randomized thrombolysis trial for stroke ever and it answers a lot of questions, such as

  • Are we safe with our (quite liberal) indications for tpa in stroke? Is the 4.5h window safe? It is, even 6h would be safe.
  • Do elder patients profit? They do! And even more than the younger ones!

The IST-3 trial is methodologically very advanced, has provided an overview of its statistical evaluation in advance. And it is actually quite readable. So go ahead and view the publication and all slides on their homepage.

Added after the fact: since I am a tpa believer (having done at least 150 thrombolyses for stroke last year) I probably don’t question the statistics enough. Make sure, you don’t fall into any trap and give the tpa agnosticians a try –  such as EM ireland.

7 thoughts on “The IST-3 trial

  1. there’s been a lot of criticism that the conclusions weren’t in keeping with the results. I’ve written some here

    as has David Newman here

    and another EM doc here

    the main concerns are primarily methodological but also with their statistical analysis

    i really appreciate the blog by the way, some great stuff.


  2. Thanks for your helpful comments, Andy. IMHO the result that not more people die if you give the drug to “bad cases” is more than enough. We all know that tpa is a balance between risk and benefit. But this trial and all the registers there are (even though they are neither randomized nor blinded) tell me that there are not too many dangers other than hemorrhage lurking, so if I concentrate one lowering that risk my patients should be fine.

    By the way:

    • that tpa leads to more dangerous swelling is not too counterintuitive – this is what is known as reperfusion edema.
    • I also find it interesting that > 4.5h should be better than 3-4.5h (if it is not a statistical artifact). One explanation might be that those patients with sufficient collaterals (hibernating penumbra w/o demarcation of stroke in CT) actually are better candidates than those that are thrombolysed despite their already apparent stroke, just because they arrived early enough.
    • Great point about reperfusion oedema – makes perfect sense , I\’ll have to
      update my post to say that.

      Agreed that the mortailty is roughly equal and I think this is reasonable proof that it helps some patients. In other words tpa kills some patients and saves others but it all balances out on a population based level. I don\’t want to stop lytics being given but I do want to know more specifically who to give it to – the imaging trials may help us with that question.

      Giving tpa is not a neutral thing on a public health level as we have to reconfigure entire emergency care systems so that stroke patients get to hospital as quickly as possible.

      If the results of IST3 are valid then there\’s a 2-3% absolute benefit to tpa which means roughly a number needed to treat of 30-50 for one pt to benefit.

  3. “Agreed that the mortailty is roughly equal and I think this is reasonable proof that it helps some patients.”
    Just think about that statement you have made for a moment.

    • sorry – should have explained that better – in IST-3 there was a big increase in mortality early with tPA (from bleeds) but this mortality difference balanced out in the long term suggesting that some people who got tPA must receive benefit for the mortality to be equal in the longer term analysis.

      the point was that tPA seems to work for some (a small number) people, and seems to harm (again, a small number) others – the key is selecting which patients are which.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.