Friday, January 17, 2014

Syncope and ST Segment Elevation. And another finding. How well does the computer interpretation perform?

A woman of approximately 50 years of age had been feeling weak and febrile, then had syncope.  EMS recorded this ECG:
Sinus Rhythm. 
The ECG computer read "ST Elevation, Anterior Injury, *****ACUTE MI*****"
The medics activated the cath lab prehospital.
Is the computer accurate?

No.  This is not the morphology of anterior STEMI, nor even of right ventricular STEMI, though there are similarities. 

There is right ventricular conduction delay (R' wave), with downsloping ST elevation and T-wave inversion (TW inversion is slight in this case).  This morphology is Brugada pattern, or at least very similar to Brugada pattern, and could actually be Brugada, though not necessarily, but it is not completely classic, as this one is

On arrival, she was febrile (39.4 degrees C), and this ECG was recorded:
The ST elevation is not as pronounced now.
What else is present?

If you look closely, it appears that the QT interval is very long in lead II (about 400 ms, with QTc of 500ms).  However, if you look at leads V2-V5 (especially V4 and V5), you see prominent U-waves.  A U-wave then probably accounts for the apparently long QT in lead II (in other words, what appears to be a QT interval is a QU interval).

The patient had a K of 3.8 mEq/L, but the Magnesium level was 1.2 mEq/L, which is slightly low.  It is well known that hypoMg causes hypoK, which of course leads to U-waves, but there is also some evidence that hypoMg alone, even in the absence of HypoK, produces prominent U-waves.  There may or may not be a causal relationship here.


1. The computer is very unreliable in diagnosing acute STEMI.  The literature would indicate that this shortcoming is usually in poor sensitivity (around 60%) but with reasonably good specificity (not many false positives).  In my experience, the computer has many false positives.  Because of the limited accuracy of the computer aided diagnosis, our protocol requires not only that the computer reads ***Acute MI***, but that the patient also have active chest pain.  Although this protocol will miss many STEMIs, it limits the false positives. 
2. The computer is very unreliable at measuring the QT interval when it appears long to the naked eye
3. The computer is very unreliable at finding U-waves. I do not ever remember a computer diagnosis mentioning U-waves.  I am not even sure if they are part of the algorithm.
4. Brugada pattern ECG is a common PseudoSTEMI pattern.  Although there is ST elevation, its morphology is completely different from ischemic ST elevation.  The T-wave inversion morphology is also completely different from ischemic T-wave inversion.


The emergency physicians immediately recognized this as Brugada pattern and de-activated the cath lab.  Closer history also revealed that she had only had pre-syncope; she remember the entire event.  There was no family history of sudden death or syncope.

She was admitted and ruled out for MI and will get cardiology follow up.  She had influenza.

Here is her ECG when not febrile:
Findings have mostly resolved.  Perhaps the prehospital ECG had some fever-induced Brugada pattern.


  1. In last ecg there is
    J point elevation in avL
    St elevation in v2
    T wave inversion in v4, V5
    Is it normal for this case?

    1. Not normal, but nondiagnostic and not worrisome.

  2. Dear Sir,

    I have a patient who showed Brugada pattern which was only unmasked and revealed by a febrile illness, and coindidentally had a family member with sudden cardiac death. However this Brugada pattern immediately resolves with resolution of fever. I would still need to consider the possiblity of Brugada syndrome right?


Recommended Resources