Thursday, May 7, 2020

Looking through a stack of ECGs for a troponin study.....

And I came across this one while reading EKGs for a high sensitivity troponin study:

what do you think?

Normally, if one sees STE in aVL with reciprocal STD in III, one would say "Acute OMI".

But I instantly knew it was not.

I immediately recognized it as normal, and classified it for the study (blinded) as "normal with normal ST Elevation and "normal ST depression."

More on this ECG:
All I know when I see the ECG for this study is that the patient has had at least 2 troponins drawn.
I don't know whether it is for chest pain, weakness, whatever.

I don't know for certain why my (our) mind immediately knows that it is normal.

I believe it is partly the flat STE, lots of upward concavity in the STE so that T-wave does not look hyperacute (hyperacute T-waves are "fat" because there is less upward concavity).

But there is terminal QRS distortion, which in a large study I just did I found to be quite good for OMI.

The ST axis is barely negative in aVF and slightly positive in II, so it is about -10 degrees, and I'm not sure what to make of that.

I sent it to Ken Grauer and he knew it was not as well.

This is what he wrote:

Hi. I’ll assume “CP” for the history.
  • Sinus rhythm. Assuming it’s an adult at least ~35, then the deep S and ST-T wave shape in V2 suggest LVH.
  • There is J-point ST elevation in I, aVL. Lead III shows some ST dep with T inversion — but given ONLY seen in lead III (aVF is flat) — I am NOT at all convinced this is necessarily acute!
  • I’d want more info, serial tracings, an old ECG, trop, etc — but my hunch is that this is not acute …
  • There are U waves (check lytes)

Why do we automatically know it is not ischemic STE/STD?

We are not entirely certain.

It remains an unfortunate fact that some aspects of the ECG are diagnosed by recognition only, like recognizing a face.

This is why I believe AI will be a force for accurate diagnosis of OMI in the future: it is good at recognizing faces, and also ECGs.  We have not yet tried to train the Cardiologs Deep Convolutional Neural Network how to recognize OMI, but I hope they will have me do that some day.

After my blinded review, I went to review the case details:

Patient was in his 30's with acute stroke.

All troponins were undetectable


Left ventricular hypertrophy concentric .
The estimated left ventricular ejection fraction is 60 %.
There is no left ventricular wall motion abnormality identified.

MY Comment by KEN GRAUER, MD (5/7/2020):
As per Dr. Smith — both of us immediately suspected the above ECG did not reflect acute changes by its appearance.
  • I share the difficulty Dr. Smith expresses in trying to explain WHY my strong hunch was that this was not acute coronary disease.
  • I thought the generous R waves as early as lead V2, in conjunction with the deep S wave and ST coving in V2 was most consistent with LVH.
  • The shape of the ST elevation in high lateral leads I and aVL is clearly upward coving — which often signals a benign repolarization change.
  • The main abnormal finding was that ST-T depression in lead III — but in isolation (ie, with a perfectly normal ST-T wave in lead II — and no more than a flat ST segment in lead aVF) — I didn’t think this was necessarily abnormal (especially given the unusual rSr’ QRS morphology pattern in lead III, in which there is predominant negativity in lead III, that is so often naturally accompanied by T wave inversion, simply because the T wave vector tends to follow closely behind the QRS vector, and that is predominantly negative in lead III ... ).

A major part of my ECG interpretation experience came from being the sole interpreter of all ambulatory tracings ordered by 35 providers in a primary care clinic over a 30-year period. I often was not given any history for the multiple tracings I had to interpret.
  • This type of ECG would be among the ~5% of tracings for which I would say to myself — “I don’t think this ECG represents acute coronary disease — but since I’m not 100% certain, I will immediately stop what I'm doing to go pull the chart (and/or call the provider) to get more history (and look to see if there is a prior tracing) — so that I can be more certain that what I'm seeing on the ECG is not an acute change.
  • BOTTOM LINE: It is often difficult to be 100% certain. Being ~90% certain (as I was after seeing this tracing) — is not “good enough” if there is a 5-10% chance that I might overlook something acute. The GOOD NEWS (as Dr. Smith has illustrated numerous times on this Blog) — is that additional measures (ie, clinical follow-up over ensuing minutes of the patient, serial ECGs, looking for a prior tracing for comparison, stat Echo during symptoms, troponin, etc) will usually allow the emergency provider to get much closer to that 99% certainty level about the patient.


  1. Thank you for another great case. I have seen very similar ECGs findings in patients who are asymptomatic during routine ECGs. I considered them as normal variants for those patients, in the absence of symptoms and in the clinical picture but was not certain. Would this be considered a less common early repolarisation pattern? I interpret around 30-50 ECGs a day for primary care in th UK and incidentally saw this pattern recently. Rapid, regular interpretation makes you realise that ECG interpretation, in it's simplest form, is just pattern recognition.

    1. @ Arron — THANKS so much for your comment and question! One KEY to optimal ECG interpretation is appreciation of the clinical history. That said, this was of no help to me — as I had NO idea if the patient in this case had new-onset cardiac symptoms or not … As to how to “classify” this tracing — I think (as per my comments above) that there are a number of things going on. The shape of the high-lateral lead (in I, aVL) ST-T wave IS perfectly consistent with a repolarization variant (concave-upward = “smiley”-configuration shape). But the rSr’ configuration in lead III, with its associated J-point, slow-downsloping ST depression is a bit unusual. Added to the mix is that QRST complex in lead V2 suggestive of LVH + nonspecific ST-T wave flattening in leads aVF and in V4,5,6 (the T wave is normally taller in these lateral chest leads). This makes for too many unique findings for a “less common early repolarization pattern”. So my approach to this tracing was to describe what I saw — and then formulate a “Gestalt overview” ( = my intuitive use of “pattern recognition”) as to whether or not the composite findings looked acute or not. Thanks again for your interest and support! — :)

  2. I sometimes see these STE patterns and interestingly the computer generally does not flag them as having ST abnormalities, either. If they have old ECGs with the same change I leave the report alone, but if it's the first ECG I usually just write something to the tune of "ST elevation; pattern favors repolarization abnormality/normal variant but can't rule out ischemia." I don't know if I should even say that second part because I don't want the GP / primary care provider to freak out when that lands on their desk, but I'm not as confident as you guys at dismissing it entirely.

    1. @ Med on the Rock — THANKS for your comment. Just to be clear — I did NOT dismiss the changes on this tracing. In my teaching for family medicine residents and primary care clinicians in practice over the past 3+ decades — I stress Systematic ECG Interpretation, in which there are 2 parts for the interpretation of EVERY tracing. PART I = Descriptive Analysis, in which I simply describe what I see (in this case, I’d write on the chart that there is high lateral lead ST elevation + rSr’ pattern with T inversion in III, flat ST in aVF + nonspecific abnormalities in lateral precordial leads + probable LVH based on the deep S in V2 with mirror-image ST-T wave that is consistent with LV “strain” in an anterior lead). PART II = Clinical Impression, in which in light of the clinical scenario — I interpret all findings noted in Descriptive Analysis, and render an “Impression” — which in this case would be something almost identical to what YOU just wrote. The 3 letters I write most often at the end of my interpretation (when I would go through a “stack” of 20-30 tracings) = “SCC” ( = Suggest Clinical Correlation) — because EVERYTHING changes IF a patient presents with worrisome, new-onset chest pain and an ECG like what we see in this case! As I stressed in the BOTTOM LINE in My Comment (above) — I usually did NOT have any history to go on when I would sit down to read that “stack” of 20-30 tracings — and an ECG like this one would leave me ~90% certain that it did not represent an acute coronary syndrome — but that remaining 5-10% uncertainty that I had would have been ENOUGH to get me to immediately STOP what I was doing to find out some history (and look for a prior ECG). Given that time is valuable — it wasn’t often that I would feel the need to immediately stop and search out more info on a given tracing — but this woud be one in which I would. Having taught primary care providers for so long — I agree 100% with you that the interpreter MUST convey the possibility (even if fairly small) of ischemia (that could be acute) — which means that IF the history was new chest pain, closer attention by the primary care provider would be immediately needed. And on those RARE instances during my 30+ years of doing this that I found something that REALLY looked acute — I’d be ON THE PHONE within seconds to find out what happened (was happening) to the patient. GREAT discussion! I hope the above clarifies my approach — :)

    2. Thanks for your reply! On reading Steve's post again, I realize he did say that he was only 90% (not 100%) sure. Great discussion, keep the cases coming! :)

  3. Interesting ECG, but I think EVERY ECG is interesting.

    I agree with both of you that while there are obvious abnormalities present, I was not struck that anything looked particularly acute from a cardiac perspective.

    Lead III is a calculated lead; it is not recorded from the patient. Only Leads I and II are recorded. Therefore, Lead III = aVF - aVL. Since the ST segment in aVF is essentially 0 mv (isoelectric), the STE in aVL will result in the downsloping ST segment in Lead III. That downsloping ST segment (to me) does not represent a reciprocal change but just a consequence of Einthoven's law. I think it was the concave STE in aVL plus this downsloping ST segment in Lead III that gave me the "non-acute" impression.

    The STE in Leads I and aVL I assumed to be early repolarization, but there is a slight problem. Early repolarization in the lateral leads is the type that in rare instances has resulted in SCD. It's not the totally benign process that it normally is in the precordial leads.

    1. Thanks JERRY. I think we are ALL on the same line of the same page on this one! Lead III could be perceived as potentially a “reciprocal change” — but to me, it was NOT quite a true “mirror-image” of the ST elevation that I saw in lead aVL + neither of the other 2 inferior leads showed any reciprocal change (ie, nothing in lead II, and no more than ST flattening in lead III) — which together with the overall benign shape of ST elevation in I and aVL was the major reason I (and I know you and Steve) suspected this was not acute. Of course, “strange things” can sometimes happen (ie, given probable LVH, perhaps the “baseline” ECG was not starting from a position of “normal” — therefore an acute change superimposed on a less-than-completely-normal baseline tracing might produce something like what we saw in this tracing, probably less than 5% of the time … ) — which is why I felt the NEED to find out about the clinical history before being more certain that nothing acute was going on. As to what we should do with the Ha├»ssaguerre on repolarization variants data — MOST of the time this is data that I wish I “didn’t know” … I long ago stopped using the term “benign early repolarization” because of his data — and prefer terms like “repolarization variant”. That said — I tended to ignore in practice that small-but-real mortality risk UNLESS the patient had a positive family history, a history of potentially malignant arrhythmias, or a history of unexplained syncope/presyncope. THANKS (as always!) for your astute comments Jerry — :)

  4. I think just cases such as this do merit great attention and deep discussion, as the presentation and the comments above clearly show here---thanks for this very intersting and difficult case!
    First I must admit that unlikely most of you, I would have not been so confident in interpreting the ECG abnormalities as reassuring and therefore not acute; as a result, my interpretation is a little different.
    While entirely agreeing on your management (immediately asking for more clinical informations, searching for prior ECGs, performing serial ECGs and troponin, stat echo etc.), while aknowledging that several ECG features are reassuring (as already underlined by dr. Smith and prof. Grauer, I would have been rather concerned even without knowing the absence of symptoms suggestive of acute coronary disease. My undestanding of the ECG abnormalities in limb leads is that if we assume the ST elevation (true, with upward concave morphology, hence reassuring) in aVL and I due to Early Repolarization, usually we should not observe any reciprocal ST depression, especially in lead III that is vectorially related to lead aVL. In addition (but to a lesser extent and less importance), if we assume the limb leads abnormalities as result of ER, then we usually see the reassutring concave STE much more frequently in infero-lateral leads rather than just in aVL (and to a lesser extent in lead I). In addition to the above concerning ECG features, should have more clinical informations provided that the patient was suffered from chest pain, I (and I suppose all of you as well!) would have been further worried, but we are told that this is not the case here.
    Finally and retrospectically, we are told the patient was admitted for stroke; this info is very important in my opinion in order to interpret the ECG abnormalities. As we all know, stroke often produces ECG abnormalities that often mimic ischemic changes: hence, unless those ST-T waves abnormalities were old, it is very likely that the STE in aVL-I, the flat ST in aVF and the STD in lead III were acute and caused by stroke.
    Great case!

    Mario Parrinello

    1. Thanks Mario for your excellent comments! This has turned out to be a SUPERB teaching case about some sophisticaed aspects of interpretation. I think we are ALL saying similar things using different words. KEY = The clinical setting is EVERYTHING. If the above ECG was obtained as a single initial tracing in a patient who presented with new-onset cardiac-sounding chest pain — then we all agree on the need to assume ACUTE until it can be proven otherwise (thru serial tracings, troponin, stat Echo, etc.). And IF we don’t initially know what the history is (which was the situation in this case) — then REGARDLESS of whether you think there is a 50% chance that findings on this ECG are acute, OR a 20% chance of being acute, OR a 5% chance of being acute — the provider NEEDS to get more information (by history, and/or ecgs, troponin, etc.) UNTIL the provider arrives at a level of comfort that the changes are not acute. And THAT was precisely what was done (ie, Steve verified that this patient had no new chest pain, and that serial troponins were negative). My “hunch” (call it my “Gestalt feeling”) was that although there clearly were ECG findings of potential concern — that these looked to me like they were not acute BUT, I did not feel comfortable stopping there without verifying that there was nothing going on clinically to suggest an acute event. Other commenters (yourself included) — felt the chance that these ECG findings were acute was higher … This is NOT to say that you or I was “right”, but rather that we each had different impressions on the likelihood that an ECG like this represented acute changes. But what I believe is MOST important — is that we ALL said the SAME thing = THIS is an ECG that could represent something acute, and additional information would be needed UNTIL we each could reach whatever our “comfort level” would be, to assure us that nothing acute was going on. Grazie ancora per il tuo meraviglioso commento! (Thanks again for your wonderful comment!) — :)


Recommended Resources