What factor determines final diagnosis of STEMI vs. NSTEMI? Is it ST Elevation? Occlusion? or Something else? What?

The answer is at the bottom, and illustrated by this case.

Written by Willy Frick

A man in his mid 30s with type 1 diabetes presented with two days of midsternal and epigastric pain, described as both "sharp" and squeezing." There was associated nausea, vomiting, and dyspnea. He said the pain was worse with supination and improved with upright posture.

What do you think?

Despite the noisy baseline, this ECG is easily diagnostic for OMI. The Queen of Hearts diagnoses OMI with 0.99 confidence (near maximum). Here is her explainability.

Here are some of the diagnostic findings:

• Very subtle hyperacute T waves (HATW) in lead I

• STE and HATW in II > aVF

• Subtle STD in V1 and V2

• STE and HATW in V5 and V6

Therefore, we have an inferior, posterior, lateral OMI. In fact, even the GE algorithm got this one (partially) right.

The emergency medicine physician documented, "His initial EKG is riddled with artifact and difficult to interpret but does not look like a STEMI." This is a very bold statement in a type 1 diabetic with very concerning sounding chest pain. One wonders why repeat ECG was not immediately performed, if artifact was felt to be a problem. The patient was treated with aspirin and a GI cocktail, which did not help the pain.

Initial hsTnI resulted at 25,994 ng/L (ref. <35 ng/L). At this point, the physician started heparin and gave nitroglycerin, which improved the pain from 7/10 to 3/10. He also obtained repeat ECG.

Once again, easily diagnostic for inferior, posterior, lateral OMI. Getting rid of the artifact increased the Queen's confidence from 0.99 to 1.0 (maximum). The ECG remains positive for STEMI by GE. The emergency physician consulted cardiology. Despite apparently hearing the above history together with two diagnostic ECGs and a troponin compatible with OMI, the cardiologist thought the ECG represented pericarditis and recommended echocardiogram.

Several hours passed with no documentation as to the reason for delay. Echocardiogram was finally performed five hours after the first diagnostic ECG. The report indicates LVEF 35-40% with "globally reduced wall motion with regional abnormalities." The cardiologist then recommended emergent transfer to a PCI center.

Upon arrival at the PCI center, he was immediately taken to cath lab. Angiogram showed thrombotic subtotal occlusion of LCx/OM1. Here is an AP caudal view before and after PCI. 

The true AV groove LCx was "jailed" by the stent and appears occluded in the post PCI image. The OM is a much larger vessel.

With the delays and recognition and transit, time from first diagnostic ECG to balloon was 15 hours and 47 minutes. This far out, the benefit of PCI is very attenuated. Troponin peaked above the upper limit of quantitation 60,000 ng/L. Echocardiogram showed LVEF 33% with akinesis of the lateral wall.

Here is the wall motion diagram. The view above is enclosed in a red box.

Final diagnosis written in the chart: NSTEMI

Discussion:

It is hard to understand how this can happen, but unfortunately the blog has innumerable similar cases. If I had to guess, I think some of the cognitive errors that may have contributed to this case are:

• The patient was young, in his mid 30s. But you are never too young to have an OMI. Even if it is not atherosclerotic, young people can have embolic OMIs.
• Here are cases of Young Women

• The ECG was perceived as having diffuse ST elevations. But it is not really diffuse -- it is inferior, posterior, and lateral. The anterior leads clearly show reciprocal change.

• The absolute degree of ST elevation (although enough to meet STEMI criteria), was still relatively small.

We also see that in the end the patient was labeled as NSTEMI, despite meeting STEMI criteria and having acute coronary occlusion. So the diagnosis does not reflect the ECG or the pathology. Instead, the diagnosis reflects how urgently he was treated.

Now, when the data are sent to the National Cardiovascular Data Registry, it will appear that the patient was treated appropriately as an NSTEMI! There is no external auditing of diagnoses selected by treating cardiologists, so missed door-to-balloon time metrics can easily be avoided by simply calling cases NSTEMI.

Smith: We usually talk about OMI that are missed because they do not meet "STEMI criteria".  But in this case, the 2nd ECG definitely meets STEMI criteria, but is nevertheless both missed AND ultimately called "NSTEMI".  We are submitting a paper showing that the most important factor in final diagnosis of STEMI vs. NSTEMI is not whether there is or is not diagnostic ST Elevation, and not whether there is or is not acute coronary occlusion, but rather whether the patient was treated in ＜ 90 minutes vs. ＞ 90 minutes.  This makes it easier to have favorable statistics of course.

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

MY Comment, by KEN GRAUER, MD (2/16/2025):  

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

Several aphorisms came to mind as I reviewed today's case. These include:

• "Junk in = Junk out ... "
•     — "Common Things are Common."
•         — "If in doubt — Repeat the ECG" (and do it soon! ).

"Junk in = Junk out ... "

Computers are wonderful instruments. But if the data we feed our computer is "junk" — then the answer the computer will give us will inevitably be "junk".

• The application of this 1st aphorism to ECG interpretation is that if your ECG is so full of artifact that assessment of ECG waveforms is unclear — then our interpretation of that ECG will be equally unclear.
• One of the most common errors I've observed over the years — is the tendency of all-too-many providers to accept "junk" on an ECG or rhythm strip when assessing the tracing of an emergency patient. While true that artifact may sometimes be difficult to eliminate (ie, in your shivering or trembling patient) — much (most) of the time the quality of your initial ECG can at least be improved by repeating the ECG with efforts to minimize artifact.
• Today's initial ECG provides a perfect example of this principle (TOP tracing that I have reproduced in Figure-1). Although we can guess at the ECG changes in ECG #1 — because of the small size of the QRS in multiple leads and the amount of baseline undulation — I found it impossible to be confident of my interpretation of this tracing from this patient with 2 days of midsternal CP (Chest Pain). Why guess when the decision of whether or not to immediately activate the cath lab hangs in the balance? Solution: Immediately repeat the ECG!

"Common Things are Common."

The consulting cardiologist in today's case further delayed the decision to perform cardiac catheterization because this cardiologist interpreted the 2 serial ECGs in today's case as consistent with acute pericarditis — despite an initial Troponin value over 25,000 ng/L. 

• Acute coronary syndromes are common in patients who present to an ED with new CP. In contrast — acute pericarditis is rare among chest pain patients who present to the ED (a point we have emphasized on numerous occasions in Dr. Smith's ECG Blog — including My Comment in the May 16, 2023 post — and the December 13, 2019 post, among others).
• KEY Point: Since it is rare to see true acute pericarditis among patients who present with new CP to an ED — that diagnosis should only be made after confidently ruling out an acute OMI (and not after an initial Troponin over 25,000 ng/L and 2 serial ECGs that should be interpreted as strongly suggestive of acute OMI until proven otherwise).
• Solution: The easiest way to avoid the misdiagnosis of acute pericarditis — is to remind yourself that most of the time the patient presenting to the ED with new CP will either have an acute coronary syndrome or "something else", but not acute pericarditis.
• P.S.: I saw no mention of having listened for a pericardial friction rub in the description of today's case. Not carefully listening for a rub when contemplating the diagnosis of acute pericarditis is not understanding that a pericardial friction rub is the single finding that could potentially be diagnostic of acute pericarditis. In my experience over 4+ decades of Attending and consulting on cases of other providers — not mentioning (and documenting in the chart) whether or not a rub was listened for in cases of suspected pericarditis is the most common oversight I've encountered. Not mentioning as a positive or negative finding whether a pericardial friction rub was or was not heard in my experience means that the provider did not listen for this.

"IF in Doubt — Repeat the ECG (and Do It Soon! )."

We are not told how much time passed in between the recording of ECG #1 and ECG #2 (beyond knowing that at least enough time passed for Troponin results to return).

• Regardless of how much time passed between the recording of these 2 tracings — it was too much time — since the artifact-laden initial ECG should have been immediately repeated by the "Junk in = Junk out" principle.

Figure-1: I've labeled the 2 ECGs in today's case.

ECG Findings in Today's CASE:

Despite the technical problems with ECG #1 — there are a number of insightful findings. As per Dr. Frick — there are subtle hyperacute T waves in multiple leads (upward arrows that I have added in).

• I enclosed lead V2 within a RED rectangle — because this is the lead that immediately caught my "eye". This lead eliminated acute pericarditis from my differential diagnosis. Among the ECG findings of acute pericarditis that I review in My Comment in the May 16, 2023 post — there should not be ST depression in lead V2 (RED arrow in this lead).
• Instead, in this patient with persistent CP and a markedly elevated Troponin — the ST depression seen in lead V2 is diagnostic of acute posterior OMI. This diagnosis is supported by the lack of any ST elevation in neighboring lead V3 (since there should normally be a slight amount of gently upward sloping ST elevation in leads V2,V3).
• The question marks and +/- label that I added in ECG #1 also negate consideration of acute pericardits, which would not show more ST elevation in lead V6 than in leads V3,V4,V5. Instead (and despite the artifact) — the unmistakable ST elevation in lead V6 confirms (until proven otherwise) that there is ongoing acute lateral OMI.
• Limb lead findings consistent with acute infero-lateral OMI are the subtly elevated ST segments in leads I,II,III,aVF — and especially the T wave inversion in lead aVL (BLUE arrow in this lead). Abnormal straightening of the ST segment takeoff in leads II and aVF is seen through the artifact — but before clinical decisions are made, this tracing should be repeated.
• Finally — there is low voltage in the limb leads. While many entities may cause low voltage — this is potentially a worrisome sign when seen in association with acute MI, as it may be an indicator of extensive infarction with myocardial "stunning" (See My Comments in the November 12, 2020 post and the January 24, 2020 post).

The Repeat ECG in Today's CASE:

Artifact was virtually eliminated in the repeat ECG. As noted — we are not told how much time passed between the recording of ECG #1 and ECG #2. There have been some serial changes between these 2 tracings. These are best appreciated by putting both ECGs side-by-side (as in Figure-1).

• QRS amplitude remains tiny in the limb leads. Considering the small size of QRS complexes in all 6 limb leads — the ST-T wave shape and the amount of ST elevation relative to the QRS is extreme.
• Why is there now ST depression in lead III? Presumably the increase in ST elevation in lateral leads I,aVL; V5,V6 from acute LCx/OM-1 occlusion attenuated the ST elevation from associated inferior infarction as it is seen in oppositely-directed lead III (lead III being the most distant of the inferior leads from lateral leads I and aVL).
• Chest lead findings in ECG #2 that further support acutely evolving posterior OMI include: i) Development of a tall R wave in lead V2; and, ii) Subtle increase in the amount of J-point ST depression in leads V1,V2,V3.

Conclusion: There is much to be learned from today's case ...