TIA

[fernandbteich]

Hello
I'd like to know the most up to date definition of TIA.
Is it "acute onset of focal neurologic signs which disappear in less than 24 hours" or must I add "with no evidence of infarction on diffusion MRI"?
A doctor told me if it's a TIA, it should appear on diffusion MRI.
To my knowledge, most TIAs are DWI-negative and if something flashes on diffusion, this is a sign that infarction happened.
What do we call these gray conditions? DWI-positive TIAs or subclinical CVAs?
Thanks

---

Members don't see this ad.

 

[typhoonegator]

Hell, there are honest-to-god strokes that don't show up on DWI if you miss them between cuts. The definition of the term TIA predates MRI. By that notion, I see no reason why a diagnosis of TIA has anything to do with the imaging. I don't like the definition you've posted, although you see it everywhere, because every migraine I've ever had would therefore qualify me for having TIAs. It needs a caveat. A TIA is a transient neurologic deficit, lasting less than 24 hours and explainable by ischemia in a cerebrovascular territory. Otherwise transient psychosis with auditory hallucinations would be a TIA.

If you could do 7T MRIs on everyone having a TIA, in the moment of their symptoms, you might very well see diffusion changes. But to use clinical imaging and require or not require there to be DWI change biases you against either long-lasting or short-lasting symptoms, so say nothing about TMB which won't show anything but is obviously relevant.

You can argue all you want about why the 24h cutoff, and whether or not imaging should be factored in. If a patient has no lasting symptoms but has DWI change, is it a stroke, or a TIA? By definition it is a TIA, but as far as I'm treating them, it's a stroke and they were just fortunate not to have any residual deficits.

Not everything that is DWI positive is infarction. Actively seizing brain foci can restrict diffusion. Oligemic brain regions can restrict diffusion reversibly, once flow is restored either through spontaneous recanalization or treatment.

 

[fernandbteich]

Thanks for the reply. So based on clinical reversibility and in order to classify patients and give them a diagnosis, I can qualify patients with transient symptoms as having a TIA (which is in line with the casual definition) regardless of their MRI result. If it happens to show a signal in diffusion sequences which explains the symptoms the patient had, I can call the TIA DWI-positive (read that 1/3 of TIA are DWI+).
I also read that half of these will show no abnormality on repeat MRI (?). A TIA has a higher chance of "appearing" as a signal on MRI if it lasts more than an hour. And the risk of a full blown CVA is higher in these patients.
Are these ideas correct? Is there any ongoing work to include imaging in the definition and triage of patients with TIAs?
Thanks

 

[typhoonegator]

I think you'll see some disagreements on what a DWI+ TIA really means to some people, and where the transition point is between that and a '"full blown CVA" as you call it. It becomes a question of who you consider to be a candidate for primary vs. secondary prevention.

For clinical trials and studies of TIA patients the phenotyping is more distinct to reduce heterogeneity -- but remember that a lot of people are still taken care of at facilities that don't/can't do a lot of acute MRIs. It's important to be able to risk-stratify patients when MRI is not immediately available. If a community hospital has only one MRI and it's doing inpatient studies all night, it's harder to justify getting a study on a patient sitting in the ED with no symptoms.

 

[Hank Rearden]

I agree with most of the ideas floating around here. There is definitely such a thing as an imaging positive TIA. When MRI came about, they realized that ~30% of the patients diagnosed with TIAs had diffusion abnormalities. I will also say that imaging positive TIA = stroke. Yes, there are rare exceptions, such as when the transient symptoms are caused by something other than stroke, but those are easy to pick out (diffusion positivity along the cortical ribbon for example in the case of seizure). Also, diffusion reversibility in the case of true ischemia is a myth. If you do repeat imaging on people presenting with suspected stroke / TIA found to have diffusion + lesion at presentation and repeat imaging at 24 hrs, 99% of them will still be positive (Freeman et al. Stroke. 2013 Jun;44(6):1629-34).

 

[neglect]

I agree with most of the ideas floating around here. There is definitely such a thing as an imaging positive TIA. When MRI came about, they realized that ~30% of the patients diagnosed with TIAs had diffusion abnormalities. I will also say that imaging positive TIA = stroke. Yes, there are rare exceptions, such as when the transient symptoms are caused by something other than stroke, but those are easy to pick out (diffusion positivity along the cortical ribbon for example in the case of seizure). Also, diffusion reversibility in the case of true ischemia is a myth. If you do repeat imaging on people presenting with suspected stroke / TIA found to have diffusion + lesion at presentation and repeat imaging at 24 hrs, 99% of them will still be positive (Freeman et al. Stroke. 2013 Jun;44(6):1629-34).

Excellent post, as were above answers. I try, and it is futile, to teach against the mindless MRI view of stroke. And it matters. If you have an MRI negative stroke, and fail to aggressively put together clinical presentation, localization, vascular imaging, secondary stroke prevention, then you're a ****ty neurologist and you're going to end up with strokes on your hands due to pure incompetence. And just the revserse. If you treat every bright DWI voxel as a stroke regardless of the clinical context: seizure, TGA, minor head injury, you will cause GI bleeds and un-needed and dangerous interventions.

TIA and stroke seamlessly merge together. One becomes the other with a very slight change in definition: based on time or tissue. And who cares? Even most trials (all I'm aware of) lump together minor strokes and TIAs. The bottom like is that they represent an ischemic insult to the brain.

 

[typhoonegator]

While the Freeman study is obviously useful in its own right and it holds up to what I've seen post-tPA in practice, I've seen a bunch of patients with mild patchy diffusion restriction (obviously not ADC black as night) who resolved after endovascular recanalization. Usually in large MCA strokes where the lenticulostriate territory is already FLAIR bright but the cortex and subcortex just has a DWI haze. I don't think it really means a whole hell of a lot, because you already know from the exam that the territory is dysfunctional, but I don't think seeing it should completely rule them out for the angio suite.