Sunday, October 1, 2023

Expert human ECG interpretation and/or the Queen of Hearts could have saved this patient's anterior wall

This is a re-post of an excellent case from 2021. See it again now, along with our new Queen of Hearts functionality. We've come a long way in 2 years! And the pace only quickens.


A man in his mid 60s with history of CAD and stents experienced sudden onset epigastric abdominal pain radiating up into his chest at home, waking him from sleep. He called EMS who brought him to the ED. He had active chest pain at the time of triage at 0137 at night, with this triage ECG:

What do you think?







I sent this ECG, without any text at all, to Dr. Smith, and he replied: "LAD OMI with low certainty. V3 is the one that is convincing." 

After his response I sent him the baseline ECG (below), still with no context at all except that this was his prior ECG:





Dr. Smith replied: "Now high certainty. By the way, the formula using QTc of 410 and STE60V3 of 3.5 was 19.4. I bet this LAD occlusion was missed."  (Also, R-wave amplitude in V4 of 12, and QRS in V2 of 16)

I replied: "You're right about both. It was LAD occlusion, and of course it was not seen until about 6 hours later." 

See side by side below:



Now lets see the Queen of Hearts interpretations for these ECGs (remember she receives ZERO clinical context, she receives nothing but the ECG waveform, and she cannot yet compare old vs. new ECG findings):


New ECG:

Overall interpretation: "OMI with High Confidence"


Baseline ECG:

Overall interpretation: "Not OMI with High Confidence"



You can see that this is an easy call for the Queen, even without any context and without the ability to compare to old. She knows the baseline is normal, and she knows the STEMI(-) OMI one is diagnostic of OMI, with the highest possible confidence. You can see that she labels the anterolateral leads V3, V4, I and aVL as the most influential leads for her decision, with V3 the most diagnostic. 

Furthermore, the blue highlighting shows the time-specific waveform features that were most influential in each lead for each interpretation. You can see that she is not only "looking" at the ST segment, but the majority of the T wave itself, and factoring in the QRS as well (just like expert humans do).


Back to the case:

Unfortunately, the ECG was not understood by the provider. Here is the EM decision making:

"The patient's EKG revealed some repolarization abnormalities but no clear signs of a STEMI. The patient's laboratory studies revealed troponin mildly elevated at 25 ng/L but liver enzymes, lipase were normal. Gallbladder ultrasound was negative for stones. Chest x-ray was normal. I ordered  morphine but he refused. I tried a GI cocktail but this did not significantly help. I felt the patient warranted admission for following his cardiac enzymes and EKG."

"ED Diagnoses:
1. Epigastric pain
2. Elevated troponin"


After hours of misdiagnosis and ongoing pain, the patient finally gets to go to the cath lab 6 hours after arrival (could have been an immediate ECG diagnosis of LAD OMI using human expertise OR the Queen), where they of course find a complete 100% thrombotic occlusion of the mid LAD, at the site of his prior LAD stent. It is stented with good angiographic result.

All subsequent troponins returned at greater than 25,000 ng/L (our lab's upper reporting limit).

Formal echo confirmed EF of 35-40%, with moderate hypokinesis of the mid-apical inferoseptal, mid-apical inferior, and mid-apical anteroseptal myocardium.


See the full post here







===================================
MY Comment, by KEN GRAUER, MD (10/1/2023):
===================================
It is often insightful to "look back" — and see from where we have come, which Dr. Meyers skillfully does with today's repost from October 5, 2021.
  • As per Dr. Meyers — ongoing development of the QOH (Queen OHearts) AI application continues to facilitate rapid and amazingly accurate OMI diagnosis. (And the utility of QOH will truly jump to an even higher level, once she learns how to compare serial tracings!).

  • That said — I focus my comments on human interpretation. With simple acceptance and application of several basic principles — there should be no way that ECG #1 gets neglected as it was in this case, in which this initial ECG was interpreted as showing, "some repolarization abnormalities, but no clear signs of stemi".

What Are These Basic Principles?
The basic principles of emergency ECG interpretation in patients who present to an ED with new symptoms include the following:
  • Respect the History. This patient in today's case was a man in his 60s with a known history of coronary disease, including prior stents. He was awakened from sleep by "sudden epigastric pain radiating to his chest" — and it was these symptoms that prompted him to call EMS. This history immediately places this patient in a high-prevalence population for having an acute event. Thus, the role of the emergency provider has to be proving that an acute event is not happening — rather than the other way around.

  • Prior tracings must be compared Side-to-Side! I do this in Figure-1. Whereas, if looked at alone — ECG #1 might be misinterpreted by a less experienced provider as "showing repolarization abnormalities" — there should be no way that you can fail to notice the change that occurs in T wave amplitudes in leads V3,V4,V5,V6 — IF you simply place both tracings Side-by-Side (ie, especially the ST-T wave amplitudes in leads V3 and V4 in the prior ECG — which were tiny compared to their size in ECG #1). In a patient with new symptoms — this amount of change in leads V3,V4 beween prior and current tracings clearly signals an acute event until proven otherwise.

  • Respect physiology. Normal R wave progression should show gradual evolution of ST-T wave morphology as one moves across the chest leads. The abrupt transition that we see in ECG #1 — from the very small, upright T wave in lead V2 — to the disproportionately larger (taller than its R wave) "bulky" ST-T wave in lead V3 — is simply not physiologic.

  • little bit of Troponin — means a lot in a case like this! The initial troponin in today's case came back "mildly elevated". But in a patient with known coronary disease — who presents with new symptoms and the above ECG changes — any amount of Troponin elevation has to be taken as indicative of an acute event until proven otherwise. (NOTE: Even if the initial Troponin would have been normal — this would not have ruled anything out, since it sometimes takes a little time before Troponin increases).

  • It's OK not to be certain from the initial ECG as to whether or not an acute event is occurring. But if you are not certain — then repeat the ECG soon and often — until you are certain. In today's case — no repeat ECG was done for hours, until finally a 2nd troponin came back markedly elevated.


BOTTOM Line: With or without assistance from QOH — this is a case that should not be missed by providers working in emergency care. It would not be missed — IF there was simple acceptance and application of a few basic principles by emergency providers.


Figure-1: To facilitate comparison — I've put the first 2 ECGs in today's case together. 





No comments:

Post a Comment

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.