Tuesday, June 4, 2019

The computer and the cardiologist called this a "Normal EKG"

This was sent by Cam Mosley, LSU-Baton Rouge Residency.

A 40-something woman presented to the ED having had “heartburn” overnight and then worsening chest pain 1 hour prior to arrival.  The patient looked ill.  Here is the ED ECG (no previous ECG was available for comparison).

Computer interpretation: Normal ECG
Cardiology overread: Normal ECG
QTc-Bazett by computer = 479 ms 

What do you think?
It is interesting that the computer measured a long QT but did not put "long QT" in the interpretation--it gave a diagnosis of "Normal EKG."  This is a systematic failure that  has been documented in the following study:

Garg A, Lehmann MH. Prolonged QT Interval Diagnosis Suppression by a Widely Used Computerized ECG Analysis System Clinical Perspective. Circ Arrhythm Electrophysiol. 2013;6(1):76-83.
Full text: https://www.ahajournals.org/doi/full/10.1161/CIRCEP.112.976803


This ECG is not at all normal!

There is high probability of inferior MI, with proportionally large T-waves in inferior leads and reciprocal ST depression in aVL.

There are what appear to be hyperacute T-waves in V1-V4.

T-waves of this size are only normal when they follow a high voltage QRS.

When this was shown to me without other information, my opinion was: Occlusion of LAD that wraps around to the inferior wall.

Let's use the 4-variable formula for LAD occlusion:
(note: Strictly speaking, to use the formula, there should be at least 1 mm of STE at the J-point, relative to the QRS onset, in at least one of leads V2-V4.  This ECG does not meet that)
But let's use it anyway:
1. Computerized QTc = 479 ms.  If you don't believe this, and measure it as shorter, let's use 400 ms.
2. R-wave amplitude in V4 = 8.5 mm
3. QRS amplitude in V2 = 8.5 mm
4. ST elevation at 60 ms after the J-point in lead V3 = 1.5 mm

Value using 400 ms QTc = 18.83 (significantly higher than 18.2 and thus indicating LAD occlusion).

This all suggests a wraparound type III LAD to the inferior wall.

Case continued

The patient continued having pain.

"The patient arrived looking like an aortic dissection patient, so CTA was done and negative."

The first Abbott Architect troponin I returned at 0.11 ng/mL (Normal: less than 0.030 ng/mL).

The pain resolved with nitroglycerin (not certain if sublingual only, or if IV).  No opiates were given.

Cardiology decided to admit the patient to medicine with a diagnosis of "NSTEMI."  She was reportedly pain free on admission.

The repeat troponin in the AM was 77.69 ng/mL (very high, very large MI), so they took her to angiography at 18 hours after ED arrival.  (This is typical for NSTEMI, even with OMI -- 25% of NSTEMI with OMI  This showed complete occlusion of an LAD that wraps around the apex to the inferior wall.

No other pre-angiography ECG was recorded.

Here is the post-reperfusion ECG:
There are now QS-waves in V2 and V3, consistent with complete infarction.


1. The first ECG was highly suspicious for LAD occlusion.  Serial ECGs were essential, but none were done.
2. No ECGs were recorded overnight
3. It is uncertain whether the patient had recurrent pain or not.
4. No ECG was recorded when after the pain resolved with Nitro.

Even if you believe the first ECG is normal, subsequent ECGs are essential.

Especially after the first troponin returned elevated, it would be essential to record another ECG.

Resolution of pain, by itself, is not reliable enough to be certain of resolution of ischemia.  One must record another ECG.  If it reverts to (truly) normal, then you can be certain that 1. the first ECG really was ischemic and 2. the ischemia has truly resolved.

OMI often does not meet STEMI criteria.  If you rely on STEMI criteria, or on the computer algorithm, many of your patients will progress to a completed MI.

This ECG showed BOTH inferior and anterior OMI and was called "Normal" by BOTH the computer and the cardiologist.

Learning Point: Record multiple ECGs

Comment by KEN GRAUER, MD (6/4/2019):
Unfortunate case of this 40-something woman who presented with new-onset chest pain, which was severe enough that a diagnosis of acute aortic dissection was considered. Although details of the case are somewhat unclear — the messages tuare emphatic:
  • Computerized ECG interpretations are notorious for their poor performance early in the course of acute OMI. One can not rely on the computerized interpretation for detecting early OMI.
  • There is no way the initial ECG in this case was normal — nor is this initial ECG at all indicative of a NSTEMI.
  • As per Dr. Smith — ordering serial ECGs should be automatic. Had just one more ECG been ordered in timely fashion on this patient — much myocardium would surely have been saved. MORAL: If not routinely ordered in timely fashion by physicians caring for a chest pain patient — Other staff should remind the treating clinician to do so. Optimal work is to work as a team.
In the hope of expanding constructive feedback from Dr. Smith’s insightful comments above — I wanted to focus on a few important points regarding ECG interpretation of the initial ECG (ECG #in Figure-1).
Figure-1: The initial ECG in the ED. For clarity — I’ve lightened the tracing, and have highlighted some key findings (See text).
KEY Concepts regarding ECG #1: As is so often the case — attention to History is of paramount importance. This patient presented with new-onset chest pain that significantly worsened in the 1 hour prior to presentation. She “looked ill” when seen in the ED. Given this History:
  • The onus of proof is on the clinician to prove this is not an acute OMI, rather than the other way around.
  • Since symptoms significantly worsened in the 1 hour prior to presentation — We should not be looking for marked ECG abnormalities! Early on in the course of evolving acute OMI — ECG changes may be subtle. This should be expected.
  • The more leads that show potentially acute ECG abnormalities — the more likely it is that an acute event is evolving. I see potentially acute ST-wave abnormalities in at least 6/12 leads in Figure-1 (leads III, aVF; lead aVL; leads V1, V2, V3).
  • Inferior leads are notorious for manifesting subtle ST-T wave changes early in the course of acute OMI — especially for leads in which QRS amplitude is small (as we’ll see momentarily for lead aVF).
  • The concept of patterns oleads” is invaluable. Appreciating this concept — I assess the various lead areas in a single glance (ie, looking at leads II, III, aVF together — at leads I and aVL together — and at several neighboring chest leads at the same time).
WHERE to Start? After assessing “the basics” for ECG #in Figure-1 (ie, the rhythm is sinus — PR & QRS intervals are normal, but the QTc is prolonged — the axis is about zero degrees — there is no chamber enlargement — a tiny wave is seen in aVL — wave progression is normal, with transition occurring between V2-to-V4) — I look next to leads in which I see definite ST-wave abnormalities:
  • There should be no doubt that the ST-T wave in lead aVis abnormal. Although the QRS complex is tiny in this lead — the J-point is clearly elevated with respect to the baseline (short horizontal YELLOW line in aVF). The magnified view of lead aVF highlights this ST elevation with associated hyperacute T wave.
  • Lead III clearly shows ST segment coving, also with a hyperacute T wave that is much broader-at-its-peak than expected.
  • Although the 3rd inferior lead (lead II ) does not show definitive changes — the scooped ST segment in lead aV(see magnified view in this lead ) shows mirror-image ST depression of the ST elevation shape that we saw in lead III. In a patient with new-onset chest pain — this is diagnostic of acute OMI until you prove otherwise.
  • In the chest leads — there should be no doubt that the ST-T wave in lead Vis elevated, with straightening of the ST segment (See the parallel RED line in V1) — plus, with a hyperacute appearance to the T wave in this lead. Normally, you just don’t see ST elevation with a prominent T wave in lead V1.
  • Support that these findings in lead V1 are real — is forthcoming from the ST-T wave appearance in leads Vand V3. Although slight ST elevation may normally be seen in these leads — the ST segment straightening (See the parallel RED lines in V2 and V3) is not a normal finding (Normally there should be gentle upsloping to the ST segment in these anterior leads — as is seen in our June 7 Case!). The base of the T waves in these leads is broader-than-expected. Taken together, the pattern” that should be appreciated in these 3 consecutive anterior chest leads (in this patient with new chest pain) — is subtle-but-real ST elevation with hyperacute waves.
  • Putting It Together: Our clinical impression of this ECG should be based on a synthesis of the ECG findings we see. In Figure-1 — potentially acute ECG findings are seen in at least 6/12 leads on this tracing. Anatomically (as per Dr. Smith) — the pattern of subtle ST elevation with hyperacute T waves in inferior + anteroseptal leads (with mirror-image reciprocal ST depression in aVL) — is perfectly consistent with acute occlusion of a “wraparound” LAD.
  • Soul-Searching Bottom Line: In a patient with new-onset worrisome symptoms — it should not have taken 18 hours for cardiac cath to be done ...
FINAL THOUGHT: We need to learn from this case. My sincere hope is that feedback was provided to the treating clinicians ...
  • Our THANKS to Dr. Cam Mosley for sharing this case, and to Dr. Smith for presenting it.


  1. Drs. Smith, Grauer and Mosley...

    Thank you for a very educational experience here. Unfortunately, it seems like this same theme has played out in several of your posts in the last few months.

    Leads III and aVF along with V2 and V3 certainly point toward a Type 3 "wrap-around" LAD occlusion. But I think there remains one more bit of information that makes this even more interesting. The ST depression in aVL (and Lead I as well) is critical to this case. I can't help but think that had they just followed (ahem!) "Jones's Rule" ("Any ST DEPRESSION on the ECG of a patient with symptoms compatible with ACS represents a RECIPROCAL change until proved otherwise."), they would have come to a resolution of this problem a lot sooner.

    But where is the occlusion? If the thrombus were proximal to D1, then the basolateral area would have been ischemic and there would have been STE in aVL as well as Lead III. While we may have had a cancellation of forces resulting in minimal to no STE in any of the limb leads, there certainly would not have been ST depression in aVL. Granted, it would THEORETICALLY be possible IF the STE in Lead III were extremely high and the STE in aVL were much less; but occlusion proximal to D1 in a wraparound LAD is going to produce more STE in aVL than in Lead III. Thus, ST depression in aVL with an occlusion proximal to D1 simply does not occur. So, the occlusion has to be distal to D1 to allow Leads I and aVL to be free to react reciprocally to the STE in III.

    But V1 - though it does not have any overtly diagnostic STE (it's still within the range of normal) - does have a genuine hyperacute T wave. But how is that possible if the occlusion is distal to D1 and S1 normally precedes D1? Simple! In THIS case, D1 is the first branch off the LAD and S1 likely follows D1. The occlusion, though DISTAL to D1, is PROXIMAL to S1, so occlusion of the septal perforators causes the ST changes in V2 and V3 while D1 is able to continue to provide circulation to the apicolateral area (V4 - V6) - with assistance from the LCx, the RCA and possibly a ramus intermedius. Of course, the continuation of the LAD that wraps around the apex is unable to supply part of the inferior wall with enough circulation, so an acute epicardial ischemia develops (i.e., STE). In the case of this unfortunate patient, infarction most assuredly followed as well.

    Jerry W. Jones MD FACEP FAAEM


Recommended Resources