A middle aged female with "heartburn" and a "normal ECG" per the computer

This is a re-posting of a Tweet by Robert Jones (@RJonesSonoEM), reproduced with permission, written by Pendell Meyers

A middle aged female with history of smoking presented to the ED with "bad heartburn."

Here is her prior baseline ECG (first), and her ED ECG (second):

Baseline:

ED ECG:

What do you think? Do you agree with the computer's interpretation of "Normal ECG"? Unfortunately the particular computer algorithm used here is not available.

This was posted on Twitter and Dr. Smith (and several others) replied that it shows OMI. Dr. Smith said "No" (meaning "No, I do not agree with the computer"). "New ST elevation in V4-V6. New distortion of S wave in V4 - OMI - distal LAD likely."

The ECG shows sinus rhythm with normal QRS and R wave progression. There is STE from V3-V6 (more than 1 mm in V4-V6), and a tiny hint in I and II. There is no STE or STD in III an aVF. Lead aVR has a bit of STD (reciprocal, as changes in lead aVR always are). There are also new very tiny Q waves in V4-6. This ECG clearly meets STEMI criteria by the way, regardless of age or gender. The T waves in V4-6 are clearly larger in area than prior, and therefore are hyperacute until proven otherwise.

Leads V4-V6 show new loss of S-wave.  When we see this in leads V2 and V3, we call this "terminal QRS distortion" (any amount of STE with BOTH: no S wave AND no J wave), but it likely has significance in any lead and requires further study.  We will study this soon with our database.

NOTICE: 

--There is STE in lead II greater than lead III. Haven't you been taught that this favors pericarditis? 

--There is also concave ("smiley face") ST segment morphology throughout. Weren't you taught that concave morphology favors pericarditis? 

--There is also new flattening of the T wave in V1 compared to prior. Weren't you taught that "new tall T wave in V1" is concerning for ischemia, and so this is the opposite? 

Rules of thumb such as these will never be sufficient to replace an expert's experience, and are not taught with enough context or expertise to explain when and why they fail. Smiley face morphology, new upright T wave in V1, STE in V1 and V2 - all of these may have some utility if you don't have the time to truly learn ECG interpretation, but are inferior compared to expert interpretation.

As always, Takotsubo stress cardiomyopathy and focal myocarditis are rare possibilities which can only be proven after a negative cath.

The ECG was repeated a few minutes later: 

The computer tries again! What do you think?

This ECG is mostly the same as the initial, except now the STE in I is slightly greater, and the ST segment morphology in III gives the subtle appearance of being depressed slightly compared to before (though it is not clearly below baseline). The Q waves are slightly larger.

Initial troponin I was undetectably low.

The patient had continued "heartburn."

The patient was taken immediately to the cath lab where an acute mid/distal LAD thrombus (TIMI flow score unknown) was found and treated. 

Peak troponin I was 37 ng/mL.  This is a high troponin (most STEMI are above 10 ng/mL for troponin I).

Here is the ECG on day 2: 

The Q waves are deeper in V3-V6, and I. II and aVF appear to have new Q waves. There is some persistent STE in V3-V6, with slight terminal T wave inversion in V3-V6 likely indicating some reperfusion of viable tissue.

From Smith's book:

Learning Points:

1. You cannot trust the computer to identify OMI, even when it reads completely "normal."

2. You cannot trust rules of thumb such as STE in II greater than III, as they could incorrectly persuade you away from the diagnosis of OMI.

3. Expert ECG interpretation can often distinguish normal variant STE from OMI from pericarditis. Normal variant STE should not have terminal QRS distortion (especially NEW QRS distortion) or pathologic Q waves.

4. Truly pathologic Q waves can be very tiny in the acute phase, when you need to notice them most. The classic descriptions of pathologic Q waves were developed from ECGs of known COMPLETED infarct, not from evolving infarct, and these criteria have limited utility in the acute setting except to confuse learners and prevent them from learning how to identify acute pathologic Q waves which can actually help the patient.

5. The first troponin in OMI is frequently undetectable, when the benefit is maximal.

6. If the ECG is equivocal or negative, bedside (Point of Care) US may be helpful if it does show a NEW wall motion abnormality (good positive predictive value), but is not terribly sensitive (inadequate negative predictive value).  Complete, bubble contrast echo is excellent: if there is no wall motion abnormality then it is very unlikely that there is a large epicardial coronary occlusion.  , especially when you or the interventionalist are not sure of the ECG findings, and likely would have shown a wall motion abnormality in this case.

7. Terminal QRS distortion likely has a role beyond just V2-V3. No ECG finding or principle applies to just one area of the myocardium - it doesn't make sense! More to come.

===================================

MY Comment by KEN GRAUER, MD (8/8/2020):

===================================

I LIKE this case by Dr. Meyers — because of the subtle changes seen on serial tracings. I focus My Comment on this evolution.

• Along the way — I’ll offer another perspective on some Learning Points in this case.

PEARL #1: In my experience — the most time-efficient way to interpret serial tracings without missing any important findings — is to pick one of the serial tracings, and to interpret this tracing completely using a systematic approach.

• After you have done so — it becomes much easier to do a lead-by-lead comparison with each of the serial ECGs that follow.

LET’S BEGIN by taking another look at the “baseline” tracing.

QUESTION: How would YOU have interpreted ECG #1 if it was the only ECG you had?

• WHAT IF this ECG-1 was the only tracing you had — and the patient was having new cardiac symptoms? What would your clinical impression be?

Figure-1: The 1st ECG shown in this case — which we were told is the “baseline” ECG (See text).

MY THOUGHTS on ECG #1 (if it was the Only Tracing I Had):

There is sinus arrhythmia (variable R-R interval — but all p waves conducting with a constant PR interval). The PR interval is normal (ie, not more than 1 large box in duration = ≤0.20 second). The QRS complex is narrow. The QTc is normal. The frontal plane axis is normal (about +25 degrees). There is no chamber enlargement.

Regarding Q-R-S-T Changes in ECG #1 — I noted the following:

• There are small and narrow Q waves in a number of leads (including tiny q waves in leads I and aVF — and in lateral chest leads V4, V5 and V6). There is a definite Q wave in lead III — that is wider and proportionately deeper (given the tiny amplitude of the r wave in this lead).
• R wave progression is normal — with transition occurring between leads V2-to-V3.

I think it’s easiest to assess ST-T wave changes by first looking at the limb leads — and then the chest leads. In the LIMB leads:

• There is subtle-but-real ST segment elevation in lead III — with a T wave that looks larger-than-it-should-be, given the tiny amplitude of the QRS complex in lead III. I didn’t think there was significant ST elevation in the other 2 inferior leads — but the T wave in lead aVF looks a bit taller-than-expected, given the small height of the R wave in this lead.
• The T wave in lead aVL is inverted. Although isolated T wave inversion in lead aVL is not necessarily abnormal when seen in association with a predominantly negative QRS complex in aVL — the T wave in lead aVL is not normally negative when the QRS is predominantly positive (as it is here).

In the CHEST leads of ECG #1:

• There is up to 1 mm of upward concavity J-point ST elevation in virtually all chest leads. There is J-point notching in leads V4-thru-V6.

Clinical IMPRESSION: I thought the ST-T waves in the chest leads of ECG #1 looked most consistent with a repolarization variant. But IF this patient was having new symptoms in association with the ECG findings we see in this tracing — it would be difficult to rule out an acute event from the appearance of the limb leads in this single ECG.

• There is after all a disproportionately large Q wave in lead III with ST elevation in this lead + T waves that could be hyperacute in leads III and aVF + potentially reciprocal ST-T wave changes in lead aVL.

It turns out that ECG #1 is this patient’s “baseline” tracing.

• KEY Point: We were not told the circumstances at the time this baseline tracing was done. This can be a critical piece of information — since our assessment might be quite different if this “baseline” tracing was obtained in the ED at the time of a previous visit for chest pain (in which case — this previous ECG might not be a “baseline” tracing after all ...).

That said — regardless of the circumstances surrounding acquisition of ECG #1 — today’s case begins with this middle-aged woman who smokes, presenting to the ED with “bad heartburn”.

• For clarity — I’ve put the initial ECG done in the ED ( = ECG #2) below ECG #1, which is the prior ECG done on this patient (See Figure-2).

QUESTION: In view of the prior tracing ( = ECG #1) — HOW would YOU interpret the initial ECG obtained in the ED?

• HINT: Since we have just interpreted ECG #1 in detail — it should be EASY to perform lead-by-lead comparison of your findings in ECG #2 with ECG #1.

Figure-2: Comparison of the prior ECG with the initial ECG done in the ED (See text).

MY THOUGHTS on ECG #2 (compared to ECG #1): The first thing to note in comparing ECG #1 with ECG #2 — is that the frontal plane axis, as well as R wave progression and QRS morphology in each of the 12 leads is fairly similar in the 2 tracings. It is always more difficult to sort out changes in ST-T wave morphology when there has been a shift in axis or an obvious change in lead placement. This is not the case here — which means that any differences noted in ST-T wave morphology are likely to be real!

Looking first at the LIMB leads:

• There is unfortunately more baseline artifact in the limb leads in ECG #2 than there was in ECG #1. The “problem” causing this artifact most likely arises from the LA (left arm) — because artifact is maximal in leads I, III and aVL. That said, despite baseline artifact — ECG #2 definitely is interpretable.
• There is now a regular sinus rhythm (there had previously been a fairly marked sinus arrhythmia in ECG #1).
• There is now subtle-but-real ST elevation in leads I and II in ECG #2, that was not present in ECG #1.
• In lead III — there is no longer ST elevation, and the size of the T wave has clearly decreased.
• In lead aVL — there is no longer T wave inversion. Instead, there is now a hint of ST elevation (proportionately consistent with the new ST elevation that we now see in leads I and II of ECG #2).
• In lead aVF — T wave amplitude has decreased (and the T wave in lead aVF no longer looks disproportionately large).
• In lead aVR — There is now some J-point ST depression, that we did not see in ECG #1.

Looking next at the CHEST leads:

• In lead V1 — the T wave is now flat in ECG #2.
• In lead V2 — there is slight decrease in T wave amplitude. In addition — there is no longer 1 mm of ST elevation, that had been seen in ECG #1.
• I see no difference in lead V3 between the 2 tracings.
• The main change is seen in lateral chest leads V4, V5 and V6. A small and narrow (probably septal) q wave had been present in these leads in ECG #1 (within the RED circles). There should be no mistaking that these small and narrow q waves have now become not only deeper, but wider Q waves in ECG #2 (within the BLUE circles).
• KEY Point: Although admittedly — the size of these lateral chest lead Q waves is still quite small — the fact that we see this change in each of the complexes in each of the 3 lateral chest leads confirms that this is a real and significant change between ECG #1 and ECG #2.
• There is also no mistaking the marked increase in the amount of ST elevation that is localized to the lateral chest leads (ie, leads V4, V5, V6).

KEY Point: Although I have written out the above specific differences between ECGs #1 and #2 in meticulous detail — in “real life”, it should take no more than 15-to-30 seconds for your eye to pick up these changes by rapidly scanning back-and-forth between these 2 tracings. Lead-by-lead comparison can be done quickly!

• After you have completed your lead-by-lead comparison — NOW it’s time to correlate these changes clinically. The patient in today’s case was having new symptoms ( = “bad heartburn” ) at the time ECG #2 was obtained. There clearly have been dynamic ST-T wave changes in the limb leads + deepening (and widening) Q waves with marked increase in the amount of ST elevation in each of the lateral chest leads.
• I interpreted flattening of the T wave in lead V1 and the loss of ST elevation in lead V2 as “reciprocal changes” to the marked increase in ST elevation in leads V4, V5 and V6.
• Putting It All Together — the history and serial ECG changes in ECG #2 all point to acute OMI, which is consistent with cath findings of mid/distal LAD occlusion.

Additional LEARNING Points: There is lots to be learned from this case. I’d add the following to the Learning Points put forth by Dr. Meyers:

• Interpreting ECGs is like telling a story. If parts of the story are taken out of context — it is easy to jump to false conclusions. Similarly — If ECGs are looked at out of context — it is easy to misinterpret them ... So, while I do find "rules of thumb" helpful — in isolation, they make up no more than a part of the "story" (ie, though still "small" — the unmistakeable increase in size of lateral chest lead Q waves in ECG #2 is clearly "significant", because in the context of Figure-2, these Q waves indicate new infarction).
• I LOVE this PEARL by Dr. Smith = “You diagnose acute pericarditis at your peril”. This Dr. Smith Pearl incorporates the clinical realities that: i) Acute pericarditis is a relatively uncommon diagnosis — especially in comparison to the frequency of acute coronary syndromes that present to an emergency setting; and, ii) Most cases that are diagnosed as “acute pericarditis” will turn out not to be acute pericarditis.
• It is therefore BEST not to diagnose acute pericarditis until you: i) Have considered and ruled out acute ischemic heart disease; ii) Know that you are dealing with a clinical situation in which acute pericarditis is likely (ie, the most common presentation for acute pericarditis is following recent viral infection — in which chest pain is pleuritic in nature, exacerbated by lying down and relieved by sitting up and leaning forward); and, iii) Have at least spent a good moment on more than a single occasion listening for a pericardial friction rub (that IF present — would confirm the diagnosis of acute pericarditis).
• By itself — ECG #2 could be consistent with acute pericarditis because there is ST elevation in a number of leads without reciprocal ST depression. That said — we should recognize that ECG #2 is not optimally consistent with acute pericarditis in today’s case because: i) The History is wrong (ie, new “bad heartburn” — No recent viral illness); ii) 3 leads show increasing Q wave size (and acute pericarditis does not produce Q waves); iii) The marked increase in ST elevation is mainly localized to the lateral chest leads — and balanced by opposing ST-T wave changes in leads V1,V2; and, iv) There is T-QRS-D in the lateral chest leads (See next bullet).
• PEARL #2 — I was happy to see Dr. Meyers suggesting that Terminal QRS Distortion (T-QRS-D) probably is helpful beyond leads V2 and V3 — which as per Dr. Meyers, are the only leads thus far for which data documents the diagnostic benefit of this finding (For brief review + illustration of T-QRS-D — SEE My Comment at the bottom of the November 14, 2019 post). What I find especially remarkable about the T-QRS-D that we see in the lateral chest leads of ECG #2 — is how S waves that descended in normal fashion below the baseline in ECG #1 (RED arrows) — have been literally “lifted up” to produce T-QRS-D in ECG #2 (BLUE arrows). The ECG picture of this type of ST elevation suggests acute OMI and not pericarditis.

As a final illustration of how EASY it is to accomplish lead-by-lead comparison of serial ECG findings — I’ve put the initial ECG from the ED on top of the final ECG shown in today’s case ( = ECG #4) that was done the following day after PCI (Figure-3). It should take no more than seconds to appreciate the serial ECG changes that have evolved:

• Q waves have deepened in multiple leads in ECG #4 (specifically in each of the inferior leads and in leads V3-thru-V6).
• Slight ST elevation is seen in each of the inferior leads.
• Transition occurs earlier in ECG #4 (the R wave becomes taller than the S wave is deep by lead V2).
• There is now less ST elevation in leads V3-thru-V6 + there is now beginning T wave inversion in each of these lateral chest leads leads.
• A normal-appearing T wave has returned in lead V1.
• BOTTOM LINE: As per Dr. Meyers — ECG #4 shows serial changes consistent with reperfusion. I thought the infero-antero-lateral lead location of evolutonary changes suggested some "wraparound" distribution to the mid/distal LAD "culprit" artery occlusion. And with this — the “story” of serial ECGs in this case is complete!

Figure-3: Comparison of the initial ECG done with ECG #4 done the following day (See text).