Sunday, January 28, 2018

Is this ECG diagnostic of coronary occlusion? Also: Inferior de Winter's T-waves on prehospital ECG??

This post was written by one of our fantastic Hennepin County Medical Center Emergency Medicine interns who is an ECG whiz, Daniel Lee.

A man is his late 50’s presents to the ED with 1 hour of post exertional chest pressure associated with diaphoresis and nausea.  He has a history of known CAD, diabetes, and dyslipidemia.

By pure clinical appearance, he looked like the textbook patient with acute MI.

This is his first ECG in the department, which I saw as it was being printed:
What do you think?

Here is his previous ECG:

This was my interpretation of the first ECG:
Sinus bradycardia with less than 1mm ST elevation in V4-V6, elevated compared to the previous ECG, suggestive of lateral MI.  Looking to the high lateral leads, instead of ST elevation that one might see in a lateral MI there is subtle ST depression in aVL less than 1mm along with new T-wave inversion.  
Leads II, III, aVF show about 0.5 mm ST elevation that is new compared to the previous ECG.  Furthermore there is a new positive T wave in lead III.  
Subtle changes, but with the history is very nearly diagnostic of acute inferior MI. 

Let's look at the 2 ED ECGs side by side:
Limb leads
Now you can see the differences more clearly

Precordial Leads

Acute ECG on the left, with slight STE in left precordial leads, 
compared to the previous ECG on the Right.

This patient had had two prehospital ECGs recorded, and these were viewed:
30 minutes after reported onset of pain:
Sinus rhythm with only a small amount of elevation in V5 and V6.  No ST elevation in the inferior leads, no T wave in III, no ST depression in the lateral leads.
T-waves are small in inferior leads.
This is a negative ECG.

10 minutes after first prehospital ECG, with continued pain:
Smith comment: This shows ST depression in multiple leads, most pronounced in V2-V4.  This is very suggestive of posterior MI.  
There is also some subtle ST depression in II, III, and aVF, and it is followed by T-waves of increased size over the previous.  

Notice how useful serial ECGs are!  

More Smith comment: it is true that ST depression (STD) due to subendocardial ischemia does not localize [it is usually diffuse ST depression, in multiple leads and not reciprocal to ST elevation in an opposite territory], this ST depression is different! The STD in V2-V4 is almost certainly reciprocal STD, reciprocal to STEMI in the posterior wall; this is evident because it is maximal in V2-V4, not in V4-V6.  

So what is the STD in II, III, and aVF?  While de Winter's T-waves have only been described in the LAD territory, there is no reason that they might not also exist in other territories.  Since this is upsloping ST depression followed by an enlarged T-wave, I believe these are Inferior de Winter's T-waves.

There is evidence that de Winter's T-waves really represent a tiny trickle of blood through the thrombotic stenosis.  There is enough limitation of flow to cause subendocardial ischemia with ST depression, and even enough limitation of flow to have hyperacute T-waves, but too little flow to have ST elevation!  I believe this based on much experience of seeing dynamic changes from STE to deWinter's and back again and correlating this with angiograms.

So this is an infero-posterior MI, with affected leads inferior (de Winter's and later STE), posterior (STD V2-V4), and lateral (STE V4-V6).

The interesting thing is that they manifest ischemia at different times:

--When there is clear ST depression of posterior MI in V2-V4, and inferior de Winter's T-waves, there is no ST elevation in V5 and V6.
--When there is STE in V4-V6, and in inferior leads, there is no ST depression in V2-V4.


The ED ECG in the context of the prehospital ECGs was indeed diagnostic of acute coronary occlusion.

Cath Results:

The cath lab was activated and co-culprit lesions were found: 99% circumflex and 95% right coronary artery (RCA).  Both were stented.

Peak troponin was (0.446 ng/mL) -- very low due to rapid reperfusion.

Subsequent echo showed no wall motion abnormality.

Learning Points

Lead aVL is incredibly useful.  New ST-depression (without LBBB or LVH) in aVL that cannot be blamed on an abnormal QRS is worrisome, and in the context of a patient with acute chest pain is almost certainly due to ischemia.  In patients with suspicion of acute MI who have any ST elevation, aVL is also a very useful lead to differentiate between pericarditis and MI.

This paper by Bischof and Smith compared inferior MI to pericarditis and found that of 154 patients with inferior STEMI, 17% of whom had less than 1 mm of STE in any inferior lead, all 154 had at least 0.25 mm ST depression in aVL.  Among 49 patients with pericarditis who had inferior ST elevation, zero of 49 had ST depression in aVL (though there are always rare exceptions such as this case).  Interestingly this study also looked at 54 “subtle” inferior MI’s, and of these 49 had some ST depression in aVL.  Thus it is not surprising that the current case of subtle inferior MI had ST depression in aVL. Moreover, T-wave inversion in aVL was also found to be 100% sensitive and 86% specific for inferior STEMI.

Here is a magnified in view of aVL side-by-side with the baseline ECG.
Presenting with STD and TWI
Old with no STD

In additions, this paper also found that all 33 patients with inferolateral MI, as manifested by STE in V5 and V6, still had ST depression in aVL.  V5 and V6 are caudal to aVL and so an inferior ST vector towards lead aVF is also slightly towards V5 and V6 but away from aVL!!  So V5 and V6 will have some ST elevation while aVL has ST depression.  Thus, even inferolateral MI has reciprocal ST depression in aVL.   

This patient functionally had an inferolateral MI given his two culprit lesions and in retrospect the elevation in leads V4-V6 were indicative that there was lateral wall involvement as well.  The lateral involvement may also explain the very subtle nature of the acute ECG findings given that much of the voltage in inferior and lateral occlusions are directly opposed to each other. 

This is the post cath ECG:
Resolution of ST depression and T wave inversion in aVL, as well as the new T wave that had been present in III.  There is some residual ST elevation in the inferolateral leads.

Summary of Learning Points:
1. New ST depression and T wave inversion in lead aVL is highly suggestive of inferior MI.
2. Even very subtle inferior MI will likely have ST-depression in aVL.  These ECGs must be scrutinized very closely!
3. Inferolateral MI will have ST depression in aVL in addition to ST elevation in V5 and V6.  
4. de Winter's T-waves may occur in any coronary distribution.


  1. Thank you again, Dr. Smith. It's very informative. But I have a question. The old ecg is normal but the new one is sinus bradycardia. May a sinus bradycardia be a suggestion of inferior STEMI? Do you have any information about this?

    1. Sinus bradycardia is certainly consistent with inferior MI, but has so many different causes (and is usually normal) that by itself it is really useless.

  2. GREAT case! I’ll add some additional subtle but important “Learning Points” to the 4 Points emphasized by Dr. Smith. Ready availability of a comparison tracing can at times be a luxury — and this is especially true here. Point #5) The BEST way to compare tracings (in my opinion) — is to first fully interpret one of the two tracings (could either be the baseline ECG or the new presentation ECG) — and then carefully compare LEAD-to-LEAD the QRST complex. Point #6) Carefully assess QRS morphology in the 2 tracings you are looking at to be sure there has been no rotation in the frontal plane and no chest lead positional change. This was KEY in this tracing — because the frontal plane QRS axis is IDENTICAL in the 1st ED ECG and in the baseline ECG from 3 years earlier. R wave progression is nearly the same (transition occurs slightly later between V3-to-V4 in the 1st ED ECG — whereas the R>S by lead V3 in the baseline done 3 years earlier). Awareness of whether there has been ANY positional change is CRITICAL for knowing how to assess whether or not there is likely to have been significant interval change. So, knowing that there has been NO limb lead positional change between these 2 ECGs — TELLS US beyond a doubt that the change to a negative T wave in lead aVL in the 1st ED tracing IS significant in this patient with new chest pain. In that context — leads I (which now shows a smaller positive T wave) — and each of the 3 inferior leads — now clearly show differences compared to the baseline of 3 years earlier. Learning Point #7) When you detect a number of subtle abnormalities (as we now have for EACH of the limb leads except aVR) — this should make you look EVEN CLOSER at the remaining leads on the tracing. And, there IS a subtle-but-real difference in ALL 6 chest leads between the 1st ED tracing and the baseline done 3 years earlier (in addition to subtle lateral chest lead ST elevation as mentioned by Dr. Smith, T waves are now more peaked in the 1st ED tracing). That this observation IS real is confirmed by the ECG done 30 minutes after onset of pain, which clearly shows even more peaked chest lead T waves. Learning Point #8) Learn to love the “Mirror Test”! Anterior chest leads show a mirror image of what you’d expect to see with acute MI. That is, if you flip over an ECG and hold it up to the light — what you see in flipped leads V1,V2,V3 will show a Q wave with ST elevation and T inversion. That’s precisely the picture we see in the 2nd black & white tracing here (done 10 minutes after the first pre-hospital ECG) — in that you have relatively tall R waves in V2,V3 with ST depression and positive T wave (which when flipped over would be a Q with ST elevation and T inversion). Taken in context with the other tracings on this patient, this 2nd B&W ECG done 10 minutes later simply reflects evolution from the B&W ECG done 30 minutes after chest pain onset (which showed prominent T waves in V2,V3 — that by the “mirror-test” look like ischemic T waves). Final Learning Point #9) Consideration of all 4 tracings we have on this patient in context with the history of new chest pain allows us to formulate a sequential evolution of ECG changes that now makes total sense. The BEST way to confirm your hypothesis about what occurred with your patient is when you can “make a cohesive story” that corresponds to the clinical history by explaining sequential evolution for each of the tracings you have available. THANKS for presenting this superb case!

  3. Great teaching case, Steve! In all my ECG courses I emphasize a diligent inspection of aVL. Failure to have a healthy respect for that one lead has resulted in a huge amount of litigation. In fact, for 40 years I've made it a personal rule to never set down an ECG without taking one last look at aVL!

    I think de Winter T waves are really an expression of REGIONAL subendocardial ischemia (ST depression with upright (usually) hyperacute T waves). This is caused by a partial occlusion of a major vessel or a complete occlusion of a branch vessel. In my own mind I feel that a regional subendocardial ischemia is caused by a partial occlusion that occurs more rapidly than the gradual "strangulation" of one or more major vessels resulting in typical circumferential subendocardial ischemia (ST depression with T wave inversion). Regional subendocardial ischemia can develop into a generalized circumferential ischemia with ST depression and T wave inversion OR the involved vessel may continue to become totally occluded creating a transmural MI that extends to the subepicardium and results in a STEMI. I think that as long as the T waves remain upright there is some localization value. Once the T waves invert, trying to localize the lesion becomes futile. While regional subendocardial ischemia has been said to only affect the LAD, I wholeheartedly agree with you that there is no plausible reason why it couldn't result from occlusion of the other two major coronary arteries (though there are probably very good reasons why it doesn't do so as frequently).

  4. I was of the opinion that 1st ED ECG is normal,but the only thing that was concerning was the subtle ST depression in aVL (? flipped T wave is normal in aVL considering the negative QRS in aVL). Overall difficult to explain this to patients with mild chest discomfort kind of symptoms and who have no prior ecgs to compare with, who then usually sign refusal consent for further observation and leave but I believe patients agree to what you tell them esp when they know you. ๐Ÿ™‚. Any suggetions to thist๐Ÿ™‚?

  5. interesting.
    its not uncommon that the medic i greet at the ambulance bay might say " ekg was just nsr doc".
    and i'm almost embarrassed to ask then to dig it out of their pockets, or sometimes to go back to the van to retrieve it. i am implying that i , while i highly respect them, am saying "i have to see that ekg", that i might be "better" at interpretation.
    but this blog, like many others, bring home the overwhelming point that every ekg, and serial ekg's when warranted, are essential.
    thanks, guys.

    1. @ Tom Fiero — I learned the "hard way" that when I was Attending, that I ALWAYS needed to see the ECG being reported on. Even when the providers were highly capable — I would catch things often that were missed ... and sometimes I'd catch truly important findings that otherwise would have been overlooked. After a while, the residents KNEW that I was going to ask for the actual ECG (and knowing that actually made THEM pay closer attention because they KNEW I'd be asking them what they saw — so THEY got better at interpreting by paying more attention!). Bottom Line — Keep asking your medics to "dig out that ECG from their pockets" — :)


DEAR READER: I have loved receiving your comments, but I am no longer able to moderate them. Since the vast majority are SPAM, I need to moderate them all. Therefore, comments will rarely be published any more. So Sorry.

Recommended Resources