Saturday, October 6, 2018

A "normal ECG" on a busy night

Written by Pendell Meyers

We walked in to an overnight shift with approximately 70 patients in the waiting room and a room full of sign-outs at midnight. At 3:55 AM during that kind of a night shift, this ECG (among many others) was brought from triage for review by my team.

We knew only that the ECG belonged to a man in his 50s with chest pain and normal vitals. No prior available.

Here is the computer interpretation:

So we have a triage-computer-normal ECG. Don't bother me with this, right?

I'm sure you can imagine the cognitive burden we had during this type of night. On first review from triage, my attending and I waffled about this ECG for a few minutes.  It is amazing what the mind will do subconsciously when it is trying to conserve dwindling energy in a stressful situation.

Then we snapped out of it and were terrified that we almost missed this. It is clearly high lateral Occlusion MI (OMI) until proven otherwise.

There is a small amount of STE in I and aVL (though slightly less than 1mm in each) with reciprocal STD in III and aVF. Lead aVL has a Q-wave which must be assumed new in this scenario. There is no clear evidence of posterior or lateral involvement, however comparison with a prior (which was unavailable) may have shown evidence of this by comparison.

Let's zoom in on I and aVL:

Lead I (top panel) has less than 1mm of STE. Lead aVL has a Q-wave followed by about 0.5mm STE with possibly terminal T-wave inversion.

We brought the patient into one of our critical care rooms and immediately got more history while recording this repeat ECG:
The STE in I has greatly diminished or entirely disappeared. aVL is also less dramatic but still abnormal. The reciprocal changes are still able to be seen. This dynamic change is proof that the first ECG was in fact representative of OMI, and hopefully means that there was some small degree of reperfusion during the 6 minutes between these two ECGs.

The patient was in his 50s with history of hypertension, diabetes, seizure disorder, and smoking, but no known coronary artery disease. The day prior to presentation (about 12 hours prior to presentation) he described sudden onset chest pain and shortness of breath while gardening in his back yard. He went inside and sat down, and the pain slowly subsided over the course of about 30 minutes. He had no further pain and went to bed that night with no complaints. He then woke up at 3:30 AM with recurrent 8/10 chest pain, substernal, non-radiating.

We activated the cath lab.

The cardiology fellow immediately came to bedside, but was unimpressed by the ECGs. He wrote in his note that "The EKG showed early repolarization in I, V2-V3 but no clear STEMI pattern."

We performed a beside US (shown below) which showed lateral wall hypokinesis as read by an US-boarded ED attending in real time.

Because I know where to look, I can see the high lateral wall not thickening or moving much in this view.

In this view I would not have been able to see a wall motion abnormality. The high lateral wall appears to be thickening.

Again, the high lateral wall in this view does appear to be thickening and contracting to me. My attending who underwent US fellowship was able to get multiple views like the first one above, and diagnosed a lateral WMA.

Unfortunately, the cardiologists asked for a point of care troponin (which you would expect to be negative at this time, less than one hour after onset of symptoms). If the patient had been "lucky," his symptoms from the prior day might have been due to ischemia prolonged and intense enough to result in small troponin increase.  No patient should have to be "lucky" by having a positive troponin to be taken seriously as a possible acute coronary occlusion. If you're waiting for troponins to turn positive, you're waiting far too long.  Occlusion is a diagnosis made on clinical and electrocardiographic and echocardiographic findings, AND ONLY TOO LATE BY TROPONIN. More importantly, the benefit of intervention for OMI must be assumed to be maximal during the early stages before the troponin rises.

Alas, it was of course negative. As was the more reliable laboratory troponin T (0.00 ng/mL).

During this time, here is another repeat ECG:
This was interpreted as worsening compared to the 3:59 ECG. Ongoing high lateral OMI.

Thankfully, the patient was eventually taken for cath. Here are the findings:
What do you see?

Fully occluded prox-mid LCX, described as "thrombotic" on report. There is also LAD disease but no full occlusion.

During intervention, lesion starting to open up.

Post PCI, full course of LCX now seen.

Troponin T peaked at 7.47 ng/mL. This is a very large MI.

See far below for data on 24 troponin T in STEMI and NSTEMI, and correlation with infarct size.

Using our OMI Progression diagram below, what do you expect to see on the post-PCI ECGs???

If successful reperfusion, you should expect to see STE and hyperacute T-waves vanish, and terminal T-wave inversions starting to form in the affected leads. Because high lateral OMI from LCX occlusion often involves the posterior wall, you might also expect to see posterior reperfusion T-waves (which show up as tall T-waves in the anterior leads).
OMI findings in high lateral leads are resolving. Q-wave in aVL remains. The T-waves in V2-V4 seem taller than before, consistent with posterior reperfusion.

Two more repeats over the next 24 hours show the progression of high lateral and posterior reperfusion:

Formal echo confirmed lateral wall motion abnormality. EF was 55%. The patient did well at least in the short term.

High Lateral OMI:

The high lateral wall is often aptly called "electrocardiographically silent." The anatomy and lead placement create very small voltage compared to the other main coronary distributions. Because everything on the ECG is proportional, this means that the ST changes of high lateral OMI will result in smaller absolute millimeter changes. Proportionally, however, they are still significant.

This difficulty results in high lateral OMI being the most commonly missed OMIs by the misguided STEMI criteria. See the proof for yourself in the literature below:

Schmitt et al. Diagnosis of acute myocardial infarction in angiographically documented occluded infarct vessel: limitations of ST-segment elevation in standard and extended ECG leads. Chest 2001
1788 patients with acute MI (diagnosed by clinical symptoms and positive CK-MB) were prospectively enrolled and all underwent emergent coronary angiography. 418 of these 1788 (23%) had acute coronary occlusion. Of the 418 patient with ACO, 29% did not meet “STEMI criteria.” The highest miss rate (50%) was recorded in patients with acute left circumflex occlusion. In the graphic below you can see the proportion of coronary occlusions which were identified by the STEMI criteria subdivided by coronary artery and presence of extended leads.

From AM et al. Acute myocardial infarction due to left circumflex artery occlusion and significance of ST-segment elevation. The American Journal of Cardiology 2010
1,500 consecutive patients with complete occlusion or near occlusion (greater than 90% stenosis with TIMI less than 3) were identified post-hoc from a prospective PCI database. Their ECGs were then re-read to determine whether they met STEMI criteria (in this study at least 1mm in 2 contiguous leads). Even with the formidable bias of knowing that these patients all had occlusion or near occlusion prior to ECG review, only 1,077 (72%) patients met STEMI criteria, while 423 (28%) did not. This is particularly remarkable considering that only 1 mm was used as criteria even for anterior MI.  The recommended value for V2 and V3 is now 1.5 mm for women, 2.0 mm for men less than age 40, and 2.5 mm for men under age 40.

Pride et al. Angiographic and clinical outcomes among patients with acute coronary syndromes presenting with isolated anterior ST-segment depressions. JACC 2010
Post hoc analysis of the TRITON-TIMI-38 randomized controlled trial comparing prasugrel with clopidogrel among ACS patients undergoing cardiac catheterization. Of 13,608 patients, 1,198 (8.8%) were identified based on isolated anterior ST segment depressions (who therefore did not meet their current STEMI criteria). 314 (26.2%) of these 1,198 patients had completely occluded culprit arteries (defined as TIMI flow 0 or 1) at the time of cardiac cath. Because only about 75% of the original 13,608 patients actually had MI, the 314 patients with occlusion actually represents closer to one-third of the group, rather than one-fourth as presented. The left circumflex artery was the most frequent culprit artery in those with occlusion (48%).

To see more high lateral OMIs, see these cases:

A 33 year old male with acute back pain radiating to the chest

"Inferior" ST depression: What is the diagnosis?

To see more OMIs that the computer called "normal" see these cases:

A 50-something woman with chest pain and 2 "normal" ECGs at triage

A middle-aged woman with chest pain and a "normal" ECG in triage

Chest Pain Diagnosed as Gastroesophageal Reflux

Another case of arrest:

Another case of arrest:

It is easy to be led astray by the computer....

Learning Points:

You must advocate for your patients, because the current paradigm and guidelines do not.

High lateral OMI is probably the easiest to miss because of the smaller voltage in these leads. Remember, everything is proportional.

"Computer normal" is not sufficient if you care about saving myocardium and saving lives. Neither is STEMI criteria. Trained humans who care about improving the care of patients with acute MI can do better. This is our profession, let's get good at it.

24 hour troponin T and correlation with infarct size.
From: Giannitsis et al.  Cardiac MRI for Quantification of Infarct Size Comparing Directly Serial Versus Single Time Point Measurements of Cardiac Troponin T.  JACC 51(3):307-14; January 22, 2008.
They studied 31 STEMI and 30 NonSTEMI.  About 2/3 of NonSTEMI have a 24 hour troponin T less than 1.0.  About 2/3 of STEMI have 24 hour troponin T greater than 1.0.  Most of the cases here with a lot of infarct mass and high troponin are STEMI, but many are NonSTEMI.
Note that our case here would have one of the highest 24 hour Troponin T levels in spite of being an electrocardiographically small MI.  

Comment by KEN GRAUER, MD (10/6/2018):
Superb case presented by Dr. Meyers! I limit my comments to reinforcing several important concepts emphasized by him. I’ll refer to the 1st ECG (done @3:53am— and the last ECG in this case (= ECG #6done ~24 hours laterFigure-1:
Figure-1: Comparison of the first and last ECGs done in this case (See text).
  • As per Dr. Meyers — it could be all-too-easy to overlook the acute high-lateral STEMI in this case. Fortunately, the “pattern recognition” data bank of Dr. Meyers prevented this. I LOVE the PEARL that stresses the magic mirror-image” relationship between Leads III and aVL. The amount of ST elevation in lead aVL of ECG #is not great. But the reason we KNOW that this is an important real finding — is that the SHAPE of the ST elevation in lead aVL is the precise mirror-image opposite of the SHAPE of the ST segment depression in lead III (I’ve superimposed within the light BLUE box the mirror-image of lead III).
  • Said another way — although the amount of ST depression in lead III is not great — we KNOW this is an important real finding — because the SHAPE of this ST depression in lead III is the precise mirror-image opposite of the SHAPE of the ST segment elevation in lead aVL (I’ve superimposed within the GREEN box the mirror-image of lead aVL).
  • With practice — these mirror-image pictures can be instantly recognized. Quite simply, whenever I suspect possible acute inferior STEMI or possible acute high-lateral STEMI — I look to see if the ST-T wave in leads III and aVL are mirror opposites  IF they are — then I assume acute OMI until proven otherwise!
  • Further support that the slight amount of ST elevation in lead aVL of ECG #1 is real comes from: iSeeing that there is ST elevation in lead I that looks acute, and proportionately seems to be more-than-that-expected with simple early repolarization; iiSeeing that the other 2 inferior leads ( = II and aVF) also show abnormal ST segments (there is straightening of the ST segment in both of these leads in ECG #1, with abrupt angulation at the junction of the ST segment and T wave); and, iiiGetting a careful History (which reveals severe 8/10 chest pain the day before, and again the next morning shortly before coming to the ED). Assessment of potentially acute ECG changes should always be extra carefully scrutinized when the history is worrisome, as it is in this case.
  • NOTE: Another reason why the amount of ST elevation in leads I and aVL may be relatively modest — is that QRS amplitude in these leads is relatively small. Especially when QRS amplitude is modest — SHAPE of the ST-T wave is more important than “amount”.
The best way to hone your ECG interpretation skills — is with rigorous clinical follow-up.
  • As per Dr. Meyers — lead aVL in ECG #6 (obtained ~24 hours latershows evolutionary changes. That is, there is now more coving of the ST segment in lead aVL (compared to what was seen in ECG #1) — and, T wave inversion is now clearly deeper.
  • The reciprocal ST-T wave changes in the inferior leads has also evolved in ECG #6, as a result of reperfusion. NOTE: IF there was any doubt about the ST segment straightening that I alluded to in the inferior leads of ECG #1 — it should now be obvious that instead of abrupt angulation — the junction between the ST segment and T wave in the inferior leads of ECG #6 is now smooth and gradual (RED arrow).
  • It would be easy to overlook evolutionary changes in the chest leads unless one does lead-to-lead comparison. Using Figure-1 to facilitate this comparison — isn’t it now obvious that the size of the peaked T waves in leads V2, V3, V4 and V5 is greater than it was in ECG #1? (Be sure to measure the height of the T waves in each tracing when doing your lead-to-lead comparison!) This increase in peaked T wave amplitude is essentially a mirror-image picture of the deepened T wave inversion that we see in lead aVL of ECG #6 — and it reflects posterior wall reperfusion.
  • Finally (Beyond-the-Core!) — Note that there is NO ST elevation any more in lead V6! Appreciation of this subtle retrospective finding tells us that the subtle ST elevation that was present in ECG #1, was yet another manifestation of acute LCx occlusion. Note also that the shape of the ST segment elevation in lead V6 in ECG #1 looks similar to the shape of ST elevation in lead I of ECG #1, yet this is different than the shape of the ST segment in V4 and V5 in ECG #1. This finding of ST elevation in lead V6 is consistent with the acute LCx occlusion that produced similar changes in high-lateral leads I and aVL.


  1. In case of NSTEMI,clinical positive, needed to be taken CKMB, CKMB mass with TnT.In UK standard ACS gives second troponin after 12 h even clinical negative patient. It is better then 1-3-6 h troponin. The best option is ECHO to see hypokinesis in borderline cases. Great case.

  2. F?/king obvious high-lateral STEMI with reciprocal ST depression!!!

    I just gave a lecture about reciprocal addition of ST changes to make the OMI diagnosis

  3. Maarten Van HemelenOctober 8, 2018 at 8:03 AM

    In addition to the high lateral infarction which should be recognizable to most skilled interpreters, I was a bit worried about the fat T waves in the first ecg - possible associated anterior wall ischemia? The angiogram still suggests to me that the distal LAD got at least some of its blood flow from LCX collaterals.

    Best regards
    Maarten Van Hemelen
    IM resident

    1. Maarten, thanks for the comment. You might be right. But if that were true then the post reperfusion ECG would have smaller T-waves, and I don't think it does.


Recommended Resources