Neuroimaging Negative Stroke – making an ASS of U and ME

30 Oct

The always excellent Emergency Medicine Literature of Note had a great post on stroke mimics and what happens when they get tPA.

It was a study trying to say that only 1.4% of their tPA patients were stroke mimics but kudos to Ryan for pointing out that a third of the cohort had -ve neuroimaging after tPA yet they were still called strokes.

He mentions that almost half of TIAs will have abnormal imaging never mind people who have a full blown stroke.

And this got me thinking about the mess we’ve gotten ourselves into with stroke and definitions and nomenclature. We’ve got to the point that we really can’t be sure what we’re talking about. Or let me put that another way –  before we had all the fancy imaging we called things strokes and TIAs and lived in blissful ignorance that we were right. Now we know we can’t be sure about what we call something but we do it anyhow!

Let me try and back that up somehow.

The WHO definition of stroke is clinical, purely clinical. There’s no scans mentioned. But we don’t practice like that. Firstly we use CT to rule out bleed, or indeed strokes too big to treat with lytics. And increasingly we’re using MR to define ischemic penumbras so we can select the folk who will do better. The WHO may have a clinical definition but we don’t.

So what could a neuroimaging negative stroke be?

  • It’s possible that people with strokes treated with tPA do so well that the area that was ischemic does not actually infarct. This is a kin to a brief occlusion causing a TIA, or in the heart to a transient LAD occlusion leaving a Wellen’s syndrome but no Q waves or infarct.
  • It’s also possible that there’s a false -ve rate to the follow-up neuroimaging.
  • It could be a stroke mimic in disguise (imagine it sang from the terraces to the tune of “god save the queen…)

But we really don’t have any hard numbers to talk about.

Some studies to show where we’re at:

1) the one Ryan talked about claimed a mimic rate of 1.4% and a neuroimaging negative stroke rate of 30%. They give the WHO definition as their justification for this.

Artto, Ville, Jukka Putaala, Daniel Strbian, Atte Meretoja, Katja Piironen, Ron Liebkind, Heli Silvennoinen, Sari Atula, Olli Häppölä, Helsinki Stroke Thrombolysis Registry Group. “Stroke Mimics and Intravenous Thrombolysis..” Annals of Emergency Medicine (October 13, 2011). PMID 22000770

2) this study of 250 tPA pts had a mimic rate of 2.8%. They don’t give us the rate of neuroimaging negative strokes but they say:

Patients, in whom an ischemic etiology was eventually the best explanation of their symptoms despite the absence of radiological proof, were labeled probable stroke. Patients in whom supportive investigations failed to establish a diagnosis of stroke or alternative diagnosis were considered as possible stroke.

Winkler, D T, F Fluri, P Fuhr, S G Wetzel, P A Lyrer, S Ruegg, and S T Engelter. “Thrombolysis in Stroke Mimics: Frequency, Clinical Characteristics, and Outcome.” Stroke; a journal of cerebral circulation 40, no. 4 (March 30, 2009): 1522–1525. PMID 19164790

3) this study showed that 23/89 (25%) had -ve MRI post tPA. They concluded 14 were TIA, 9 were mimics. They did not conclude that they were neuroimaging -ve strokes.

Giraldo, Elias A, Aisha Khalid, and Ramin Zand. “Safety of Intravenous Thrombolysis within 4.5 h of symptom onset in patients with negative post-treatment stroke imaging for cerebral infarction..” Neurocritical care 15, no. 1 (August 2011): 76–79. PMID 21394544

4) this study found 21% (of 500) had -ve scans post lysis. They said that 14% were mimics and the other 7% were neuroimaging negative strokes.

Chernyshev, O Y, S Martin-Schild, K C Albright, A Barreto, V Misra, I Acosta, J C Grotta, and S I Savitz. “Safety of tPA in stroke mimics and neuroimaging-negative cerebral ischemia..” Neurology 74, no. 17 (April 27, 2010): 1340–1345. PMID 20335564

After these four, I think I’m happy to say that 20-30% of the time when you give tPA to someone who looks like a stroke, they’ll have a negative follow-up scan. What I can’t say is whether or not they’ve had a stroke or not.

And I know the neurologists are good, they’re better than me at calling a stroke no doubt, but I’m suspicious that they’re not right as often as they claim.

UPDATE: Aaron has a great post on stroke mimics in general from a few days ago that is well worth a read.

UPDATE April 2013: Ryan has written another nice post confirming that the idea of the “neuroimaging negative stroke” is  likely not common, it it’s a thing at all

3 Replies to “Neuroimaging Negative Stroke – making an ASS of U and ME

  1. I personally think the mimic rate is far higher than the reported rate. I think this is particularly true in community hospitals without access to neurologists, fast CT-angiography or MRI. In Canada as emergency physicians in community hospitals we are expected to make the thrombolysis decision without neurology support, and without CTA or MRI to clarify potential stroke mimics.

    I put a post about this on my blog at : http://www.emergsource.com/?p=446

    I honestly think that the literature around stroke mimic is fatally flawed as most of this literature comes from centres with neurology stroke teams, while most of the mimic risk (those whose mimics will bleed such as neoplasm, subacute stroke, small difficult to see bleeds, and brain abscesses) occurs in community hospitals where the thrombolysis decision is made the emergency physician without access to adequate neurology consultation, and without access to more detailed imaging when the issue of a potentially dangerous mimic is in play…..

    Aaron

    • Hi Aaron

      Cheers for the comment, Stroke is a whole different beast to diagnose than STEMI (back when STEMIs got lysed…) and I know I feel happier when it’s someone else making the decision!

      I’ve added a link to your post in the post above

      Andy

Leave a Reply to Aaron Cancel reply

Your email address will not be published.

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