Everyone sees ST depression, but what does it mean?

Written by Willy Frick

A man in his 70s with hypertension and type 2 diabetes mellitus presented with chest pain which awoke him from sleep around midnight. He described it as substernal, non-radiating, 7 out of 10 intensity. His ECG obtained around 4 AM is shown.

ECG 1

What do you think?

I texted this ECG to Dr. Smith without any information and he immediately replied: "combination of precordial swirl and South African Flag sign, with Swirl predominating.  Occlusion proximal to septal perforator AND first Diagonal."

The Queen of Hearts easily diagnoses OMI. Her confidence is 0.99, almost maximal. In particular, we see:

• Precordial swirl

• STE and hyperacute T waves (HATW) in V1, V2
• Reciprocal STD in the lateral precordium, V4-6

• (Most of) South African Flag

• STE and HATW in aVL and V2 (less apparent in lead I)
• Reciprocal STD in III

This is diagnostic for LAD OMI involving the septal perforators (precordial swirl) and major diagonal (South African Flag). This patient should be taken for immediate angiogram. 

The ECG was interpreted as showing "mild ST depressions in leads II, III, aVF, V4, V5, V6." The patient received aspirin 325 mg and nitroglycerin 0.4 mg sublingual. According to documentation, his pain resolved (but there was no short interval repeat ECG to assure normalization). First troponin I (cTnI) was 0.049 ng/mL (ref: < 0.033 ng/mL). 

The next update in the chart is at 8:10 AM, when repeat cTnI was 3.788 ng/mL, already a sizable infarct. After this resulted, repeat ECG was obtained:

ECG 2

Compared to the first, there is obvious improvement, but it is not normal and one must wonder whether there could be ongoing ischemia. The admitting service described the ECG as having "ST depressions in II, III, aVF, V4-V6," and documented that his pain was significantly improved to 2/10 and resolved during the interview.

PAUSE

Two thoughts:

First, there either is ischemic chest pain, or there is not. "Nearly gone" means unresolved. To use an analogy, would you be happy to drink water with nearly all the sewage filtered out?

Second, when evaluating a patient with ischemic chest pain, it is the clinician's job to elicit symptoms. The specific questions we ask encourage various responses. If you tell the patient "This nitro should help with the pain...Are you feeling better?" they might give you reassurance.

When I treat patients with suspected ACS, I tell them: "Your symptoms are critically important to me. If you feel absolutely normal, we can safely wait until the morning. On the other hand, if you have ANY chest discomfort [or whatever symptom the patient was experiencing] I need to know immediately. Please tell me or tell your nurse, because that would be an emergency."

Back to the case:

The patient was started on continuous heparin infusion and taken for echocardiogram. Apical views are shown below, first 4 chamber, then 2 chamber, then 3 chamber.

The apex is akinetic. It almost looks like takotsubo except we know from ECG and history that it is LAD OMI. The LVEF is ~45% Around this time, the patient was evaluated by the cardiology consult service. The consult note describes the ECG as showing "STD in the inferolateral leads," making this the third interpretation to overlook diagnostic STE and HATW in V1, V2, and aVL.

The patient was taken for cath. I have included the RAO cranial and RAO caudal projections here. Have a look at them and see if you can figure out the diagnosis before watching the video that follows where I explain. As always, you can read more in the angiography guide.

RAO cranial

RAO caudal

Narrated Video Explanation

Color scheme for cath films by Willy Frick.

The angiogram shows exactly what the ECG predicted: LAD occlusion involving the septal perforators and the major diagonal vessel. This is considered a type 2 MI since it is not due to atherosclerotic plaque disruption.

At around 12:30 PM, the patient received intracoronary nitrogylcerin with resolution of vasospasm. Repeat cTnI was still rising when last checked, 6.607 ng/mL.

Discussion:

Cases like this should infuse us with humility. Documentation said the patient's chest pain resolved after NTG, but the next note said he had only "2/10 pain." By the time he got to cath 8 hours after his first diagnostic ECG, his LAD remained severely spastic with very poor flow. If the first ECG had been a STEMI, he would likely have been treated emergently. But instead, it was the red-headed stepchild, NSTEMI. And this turned into another case of supervised, in-hospital anterior wall infarction. Cases like this make it hard to feel comfortable delaying intervention due to symptom improvement. If the patient comes in with ACS symptoms and ECG showing OMI -- just get them to the lab. 

This also brings up the general topic of non-atherosclerotic causes of myocardial ischemia and infarction. JACC published an elegant trial called CorMicA, in which investigators took patients with angina and no obstructive coronary disease (also called ANOCA or angina with non-obstructive coronary arteries) and performed comprehensive coronary evaluation with vasospasm provocation testing and microvascular evaluation. The patients were then randomized to routine care (in which the treating clinician was blinded to the results of the comprehensive assessment) vs stratified medical therapy according to the results of the invasive diagnostic procedure.

CorMicA, JACC 2018

Patients in the intervention arm had significant improvement in symptoms (the primary endpoint) and quality of life. In addition, patients in the control arm were 12.5 times more likely to have a missed diagnosis!

Patients with vasospasm should be strongly supported in their efforts to quit smoking which is a known risk factor for vasospasm. In addition, they should receive calcium channel blockers and long-acting nitrates to prevent recurrent spasm.

Learning points:

• It is your job to ELICIT symptoms of refractory ischemia. Delay intervention at your own peril.

• "Much better" is not the same as "completely resolved"

• Coronary vasospasm classically occurs in the middle of the night or early morning

• Help patients quit smoking, and treat with calcium channel blockers and long acting nitrates

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

MY Comment, by KEN GRAUER, MD (3/30/2025):

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

As per Dr. Frick — "Today's case should infuse us with humility" — as there are many aspects of this patient's care that could have (and should have) been improved. Filling us with humility is the frequency with which cardiac catheterization (as it is currently performed in most institutions) fails to identify all-too-many patients with a significant component of either microvascular and/or vasospastic angina (Results cited in Dr. Frick's above discussion, taken from the CorMicA Trial by Ford et al: JACC, 2018).

• The problem is not simple — as these non-atherosclerotic causes of CP (Chest Pain) run the gamut of symptomatology — and may be seen either in patients with pure coronary spasm — or — in patients with variable degrees of underlying atherosclerotic disease, upon which either a microvascular or vasospastic component is superimposed in any one or more of the major coronary branches (Matta et al — J Interv Cardiol, 2020 — and — Slavich and Patel — Int J Cardiol Heart Vasc, 2016).
• And then there is the issue of MINOCA (Acute MI in patients with Non-Obstructive Coronary Arteries) — which is estimated to occur in ~10% of patients initially diagnosed as having a STEMI or NSTEMI — in which microvascular and vasospastic angina are just 2 of the potential causes of an MI for which there is no evidence for a "culprit" vessel (See My Comments in the June 5, 2024 post — and the December 19, 2023 post).

The above said — I focus My Comment today on aspects associated with this patient's presentation and his initial ECG.

• To facilitate comparison of today's initial ECG with the repeat ECG — I've put both tracings side-by-side in Figure-1.

QUESTION:

• How to decide IF (and When) cardiac cath is indicated?

Figure-1: Comparison between the first 2 ECGs in today's case.

A Simplified Approach: When to Cath?

Cardiac cath was not performed in today's case for a full 8 hours after the initial ECG was recorded. As emphasized by Dr. Frick — this delay was entirely unnecessary, since the initial ECG was diagnostic of acute LAD occlusion until proven otherwise!

• Identification of the primary problem in today's case ( = coronary spasm) could not be diagnosed until cardiac cath was performed. But the cause of today's symptoms and ECG changes is not the main issue here. Instead — the principal issue is whether and if so, when cardiac cath should be performed in a patient such as the 70-something year old man in today's case who presented to the ED for severe new-onset CP that awakened him from a sound sleep at 4:00 am.

The Answer to the above clinical question should be simple:

• IF the initial ECG of a patient with new-onset CP is clearly abnormal (and clearly suggestive of an ongoing acute cardiac event) — then nothing else can possibly negate the indication for prompt cardiac catheterization.
• By "clearly abnormal" — this does not mean an ECG that satisfies millimeter-based criteria for a STEMI. This is because at least 30% of acute OMIs do not initially meet STEMI criteria — and many of these acute coronary occlusions that eventually do satisfy STEMI criteria, only do so after many hours of delay (Ricci, Smith, et al — Ann Emerg Med 85(4): 330-340, 2024).
• Although an elevated initial Troponin level may confirm acute infarction in progress — this initial Troponin level is not a part of the initial decision-making process IF the initial ECG is clearly abnormal. This is because the initial Troponin level may be normal in as many as 25% of patients who present with a STEMI (Wereski, Smith et al — JAMA Cardio 5(11):1302-1304, 2020). Therefore, if the initial ECG is clearly abnormal — the decision of whether or not to perform prompt cath remains the same regardless of whether or not the initial Troponin is normal or elevated (since a normal initial Troponin in no way rules out an acute event).
• And although repeat ECGs may demonstrate ongoing evolution of acute infarction — Waiting so that you can repeat the ECG should not be a part of the initial decision-making process IF the initial ECG is clearly abnormal. This is because no matter what a repeat ECG shows (ie, worsening — improvement — or no change at all) — IF in a patient with new-onset CP the initial ECG is clearly abnormal — Waiting only delays the indication for prompt cath that has already been established (and waiting only means that more potentially salvageable myocardium will needlessly be lost).

The Initial ECG in Today's CASE:

As per Dr. Frick — today's initial ECG is clearly abnormal (TOP tracing in Figure-1):

• In today's patient who presents with severe new-onset CP — my "eye" was immediately drawn to the 2 leads within the RED rectangles. In view of the small S wave in Lead V1 (ie, There is no LVH) — there simply should not be the ST segment straightening and the amount of J-point ST elevation that we see in ECG #1.
• In this context — the T wave in neighboring lead V2 is also disproportionately tall. An additional subtle-but-real finding is the presence of a small initial q wave that precedes the small r wave in lead V2 (within the dotted BLUE circle).
• The BLUE arrow in neighboring lead V3 highlights loss of the small amount of gently upsloping ST elevation that is normally seen in lead V3. This is followed by 1-to-2+ mm of flat or downsloping ST depression in lateral chest leads V4,V5,V6.
• The other RED rectangle highlights lead aVL — that manifests ST segment coving with >1 mm of ST elevation. This ST elevation in lead aVL is complemented by clearly abnormal reciprocal ST depression in each of the inferior leads.
• BOTTOM Line: Nothing else is needed in today's case to establish the need for prompt cath as soon as this can be accomplished.

The Repeat ECG in Today's CASE:

4 hours later — a repeat ECG was obtained (ECG #2 in Figure-1). That said — multiple concerns relate to this repeat ECG.

• The patient was given sublingual NTG after ECG #1 was recorded — which chart records indicate resulted in "resolution of the patient's CP". As soon as the patient's CP was resolved — the repeat ECG should have been recorded. Instead — no ECG was repeated until 4 hours later, and only then because an elevated Troponin value returned from the lab.
• With acute ongoing infarction — the initial ECG should be repeated within no more than 15-20 minutes after the initial ECG.
• As per Dr. Frick — 2/10 CP (as was noted at the time ECG #2 was recorded) is not the same as 0/10 CP.
• Rather than description of ECG #2 as having "ST depression in leads II,III,aVF; V4-V6" as was written by the admitting service — there have been "dynamic" ECG changes that become obvious by the side-by-side comparison of ECG #1 and ECG #2 afforded in Figure-1 (ie, in the form of deflation of the ST elevation in leads aVL and V1, and deflation of the hyperacute T waves in leads V2,V3). There has also been evolution of reperfusion T waves in leads III,aVF; and in leads V3-thru-V6.
• BOTTOM Line: Demonstration of these "dynamic" ST-T wave changes in association with the reduction in CP severity — confirms an ongoing acute event until proven otherwise. That said — the indication for prompt cath had already been firmly established 4 hours earlier at the time the initial ECG was done.

Even in retrospect, no one could see it.

This was sent to me by a former resident who is outstanding at reading ECGs for OMI.

"Hi Steve wonder what you think of this ecg in a 60 yo woman w cp, known CAD"

Presentation ECG (ECG 1):

Here is her previous from one week prior when she presented with heart failure and trops were "negative" (ECG 2):

My response: "They both look like active ischemia.  The top one (ECG 1) looks like "precordial swirl", which is LAD Occlusion.   The previous ECG also shows active ischemia."

The Queen of Hearts diagnosed "Subendocardial Ischemia"

More explanation: Both have some lateral precordial ST depression which looks like subendocardial ischemia.  However, the presentation ECG (ECG 1) has large, bulky T-waves.  By itself, it is very concerning for LAD Occlusion with Precordial Swirl morphology, but especially when comparing to the old one, those features of hyperacute T-waves really stand out.

He stated later that he gave her 1 sublingual NTG and her pain went down to 1/10.  He did not repeat an ECG.

His response:

"Yeah, that’s what I was afraid of. This is a retrospective ask because with the most recent ECG (ECG 1), I admitted her with new Chest Pain and a modest troponin elevation that I thought was progression of bad multi-vessel disease but she ultimately proved to have acute LAD occlusion that didn’t go to Cath for almost 24 hrs and I’m kicking myself.

Smith: In other words, he saw it in retrospect.

"That is what I was worried about in retrospect. Big MI. They stopped trending trop at 5000 ng/L of hs trop I. EF 20% down from 50%."

"And… they hit the carotid trying to place an Impella, caused a big hematoma where she had precipitous airway occlusion. Just awful all around. I’ll check for updates and other ekgs when I log in later today."

Post cath ekg here. She’s in impella dependent cardiogenic shock w new renal failure. 

Now there is a new intraventricular conduction defect (IVCD) that is very much like true Left Bundle (LBBB).

It is Smith Modified Sgarbossa Negative, but that is because the damage is done.  

Smith: "Sorry about all that."

Former Resident: "Yeah. Me too. Appreciate your feedback. I couldn’t get anyone else to agree with me on the OMI findings even in retrospect and it was driving me crazy."

Smith: In other words, no one else could see the OMI, even in retrospect.  I could see it because Pendell and I have described and coined the term "Precordial Swirl".  We have a paper In Press at the Journal of Electrocardiology that defines it based on a lot of data.  The Queen of Hearts did not diagnose OMI, but she did see ischemia.  

Refractory ischemia needs the cath lab emergently.  

When can you say that the ischemia has resolved and the patient does  not need emergent angiography?  

Both:

1) Pain is resolved completely

2) ST depression is competely resolved

In this case, the pain was still 1/10.  He told me that he did not see any inpatient notes about the presence or absence of pain.  He did not get a repeat ECG after the NTG.

SUMMARY:  This devastating OMI could have been reperfused quickly if the presence of refractory pain had been acted upon, regardless of the ECG.  Troponin was elevated and pain was refractory.  Why was it not acted upon?  I can't say I know for certain.  This former resident is one of the smartest and most diligent I have ever known, but the pressure to not activate the cath lab because of a possible false positive activation is intense.

It is better to have a false positive activation than it is to have a disaster like this, and all members of the team should be encouraged to support taking the patient for angiogram when there is any doubt!!

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

MY Comment, by KEN GRAUER, MD (3/27/2025):

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

Today’s case by Dr. Smith is subtle — and EASY to overlook because of how abnormal this patient's previous ECG is.

• This 60-year old woman clearly presented to the ED as a higher-risk patient — given her history of known coronary disease, now with new chest pain.
• To facilitate comparison in Figure-1 — I've put today's initial ECG (TOP tracing) — together with this patient's previous ECG from 1 week earlier (LOWER tracing). Of note — Troponin was normal at the time ECG #2 was recorded, so there was no acute infarction in association with this tracing.

PEARL: The KEY for assessing serial tracings is to systematically compare ECG findings by going lead-by-lead. It does not matter if you look first at the previous tracing or at the most current tracing — as long as you then systematically compare one lead in one of the tracings with the same lead in the next tracing — and then repeat this process for each of the 12 leads in both tracings.

• The Clinical Reality: If you do not keep both tracings close to each other as you go lead-by-lead comparing ECG findings — You will miss subtle abnormalities. I believe this was the problem in today's case.

The previous ECG is very abnormal ...

Let's begin with ECG #2 — which is the previous tracing that was recorded ~1 week earlier at the time of a heart failure exacerbation. 

• The rhythm in ECG #2 is sinus at ~75/minute, with several early beats. The PR interval is normal — the QRS is narrow — and the QTc may be borderline prolonged.
• There is marked left axis, consistent with LAHB (Left Anterior HemiBlock).
• Voltage criteria for LVH are satisfied (R wave in lead aVL ≥12 mm).
• There is poor R wave progression with delayed transition (The R wave does not get taller than the S wave is deep until between leads V5-to-V6).
• Regarding ST-T wave findings — There is fairly diffuse ST segment flattening with slight depression in multiple leads — with ST coving and slight elevation in lead aVR.
• Impression: Although the downsloping ST depression in high lateral leads I and aVL could reflect LV "strain" — the overall picture in this tracing is that of DSE (Diffuse Subendocardial Ischemia). That said — We are told that Troponins were negative at the time this ECG was recorded. So, while ECG #2 may be indicative of significant (even multi-vessel) coronary disease — it is not diagnostic of acute coronary occlusion.

Figure-1: Comparison between the first 2 ECGs in today's case.

Why an Acute Event was Not Diagnosed in ECG #1:

Note that both the frontal plane axis and precordial lead QRS morphology for the 2 tracings in Figure-1 are very similar (with the exception of predominant positivity for the QRS in lead V6 of ECG #2 — whereas transition never occurs in ECG #1). As a result of this similarity in axis and overall QRS morphology — any slight changes in ST-T wave morphology are likely to be "real".

• What differences do YOU see between the 2 tracings?

ANSWER:

The heart rate is somewhat faster in ECG #1. Once again there are some early beats that appear to be supraventricular.

• The most concerning difference in ST-T wave morphology between ECG #1 and ECG #2 — is seen within the RED rectangle. In this patient with new CP (Chest Pain) — Isn't it much easier to appreciate the taller T waves in leads V1,V2,V3 with these 2 ECGs placed next to each other?

Other changes are more subtle.

• I thought the ST-T waves in each of the 4 leads highlighted by BLUE arrows looked more acute (ie, with more angled and deeper ST depression).
• Lead aVR shows more ST elevation.

Impression: As per Dr. Smith — New ST elevation in anterior leads V1,V2 — with more ST depression in lateral chest lead V6 — is consistent with precordial swirl (See the October 15, 2022 post — for multiple examples of Swirl — and My Comment at the bottom of the page for qualitative summary of ST-T wave changes with this entity). The increased ST depression in multiple leads in ECG #2, with more ST elevation in aVR — is consistent with severe underlying coronary disease with worsening ischemia.

• Isn't it much easier to recognize subtle-but-important differences between 2 serial tracings by putting both ECGs next to each other, and then going lead-by-lead in your assessment? 

 

 

Can you localize the culprit lesion on angiogram without taking ECG findings into account?

Written by Willy Frick

A woman in her 60s with very severe hyperlipidemia (LDL >200 mg/dL) presented with acute onset chest pain. Her symptoms began while getting off the bus. She described the pain as moderate in severity, and said it had come and gone several times over the next few hours before ultimately resolving. On day of presentation, the pain returned and was very severe with associated dyspnea and vomiting. Given this, the pre-test probability of OMI is very high. High enough that a bland ECG is not immediately reassuring.

ECG 1

What do you think?

To me, this ECG is not diagnostic. The inferior T waves are a bit generous, but they could be baseline. There is low voltage in the precordium which always makes reading ischemia harder. There is sort of an eerie flatness to the ST segments in V2-3. But if you call every ECG that looks like this OMI, you are going to have a lot of false positives. In the ER, the patient received nitroglycerin paste. Pain was unresolved, so she received a nitroglycerin tablet and IV morphine with resolution of her pain.

Smith: it is highly suspicious for inferior OMI, but not 100% diagnostic. It has some morphologic features of LVH, but not the voltage. The Queen of Hearts does not diagnose OMI.  

In ACS, chest pain is the warning sign of ongoing ischemia. I have said before, treating angina with morphine and continuing non-emergent management is like taking the batteries out of an actively alarming smoke detector during a house fire and going back to sleep.

Smith: As Willy says, and as we've said many times before, morphine will resolve pain without resolving ischemia.  So ongoing symptoms must be assumed if opiates have been given.  ONLY give opiates if the pain is intolerable or you will activate the cath lab at the first objective evidence of coronary ischemia.

No further ECGs were obtained for the rest of the day. The first high sensitivity troponin I (hsTnI) was 745 ng/L.  

Smith: At this point, it is acute MI with presumed persistent refractory symptoms and the cath lab should be activated.

Case continued

Alas, this simple rule was ignored.

Subsequent troponins climbed: 588 ng/L, and 4189 ng/L (reference: < 14 ng/L). The following tracing is taken from telemetry around midnight:

From chart review, it looks as if VF persisted for 20 minutes before successful termination. Documentation indicates that the patient was shocked 4 times (with no comment on energy level) and received amiodarone 300 mg IV and magnesium 2 g IV. Getting patients out of VT/VF is the number one priority. A shock can fail for two reasons, and this can be addressed with two different approaches:

1. The shock terminates VT/VF, but then it comes back immediately
2. The shock never terminates VT/VF, even for a second

In the first case, the shock did its job but there are other factors promoting VT/VF such as unrevascularized ischemia. In this case, you can give anti-arrhythmics and address the underlying cause (OMI). In the second case, the patient never converted meaning the shock did not do its job at all. In this case, you should get a second defibrillator and perform double sequential external defibrillation (DSED). Simply attach a second defibrillator as shown in the diagram below and deliver max shocks from both devices simultaneously.

Smith: often, providers think that the above rhythm is torsade (or polymorphic VT due to long QT).  But torsades and ventricular fibrillation cannot be distinguished by the monitor.  They can only be distinguished by:

1) if there are pulses, then it is not VF

2) if it spontaneously resolves, then it is not VF, with rare exceptions

DOSE VF, New England Journal 2022

Remember that in the trial DOSE VF, pre-hospital use of DSED increased the likelihood of survival to hospital discharge by 17.1% in absolute terms! That's a number needed to treat of just under 6 and it costs almost nothing.

Back to the case:

Repeat ECG after return of spontaneous circulation is shown.

ECG 2

We now have clear inferior, posterior, and lateral OMI. The patient was taken to lab for coronary angiography. All the angiograms (and much more) are available on the angiography guide if you wish you read in more detail.  For this post, I have included the most important angiogram followed by my narrated explanation. The rest of the angiograms are here.

And here is my narrated explanation:

The patient had intervention to his proximal LAD. No intervention was performed on the LCx. The report says that there was "occlusion of a smaller lateral branch." Repeat ECG twelve hours later is shown below.

ECG 3

This ECG shows persistent posterolateral occlusion. The inferior infarct appears nearly complete (prominent Q-waves with T-wave inversion).  Thus, it appears as if the top ECG was truly an inferior OMI. There is no evidence of reperfusion, only progression of untreated infarct. Serial hsTnI beginning at the time of the arrest were 9652 ng/L, 12,747 ng/L, and 21,112 ng/L. None further checked. Echocardiogram was technically difficult but showed preserved ejection fraction and mid to distal inferolateral hypokinesis.

The patient remained intubated for several days following this. Even after extubation, she was amnestic for the events leading up to hospitalization, so it is unknown whether she was experiencing persistent chest pain while intubated after stent to the LAD. She was awake and following commands, but no more specific details about her neurologic exam are available.

Discussion:

This case very likely represents another case of wrong vessel PCI. Many similar cases have been featured on the blog. I shared this case with several people, and no one was completely certain which vessel was the culprit.

I think it is likely the OM for the reasons I mentioned in the narrated video, however it is worth wondering whether a vessel that small could produce ECG changes that extensive. And it is definitely possible that a more proximal LCx lesion ruptured and produced distal embolism. Some others worried about the RCA as a possible culprit.

Uncertainty about culprit vessel is very common. In fact, in this elegant study by Heitner et al., investigators found that among patients receiving revascularization, 27% received revascularization solely to non-culprit coronary arteries!

A 2017 trial named CULPRIT SHOCK found that in patients with cardiogenic shock, a strategy of culprit vessel PCI only was associated with better outcomes than immediate multivessel PCI. After CULPRIT SHOCK, many shied away from multivessel PCI in the acute setting. However, we must bear in mind two things when translating this evidence to clinical practice. First, these were patients in shock. And second, this presupposes clear identification of culprit (an inclusion criterion for that trial). Many patients may not have a clear culprit.

Other trials have found a benefit to revascularizing non-culprit lesions when a patient presents with ACS, the largest of which was COMPLETE, and more recently MULTISTARS AMI. Although I believe the results of these trials, I always had trouble squaring these findings with the results of COURAGE, BARI 2D, ISCHEMIA, REVIVED-BCIS2 which showed that stenting stable lesions does not result in improvements in major adverse cardiovascular events (MACE). And in a study like COMPLETE or MULTISTARS AMI, you are stenting an acute plaque rupture (the culprit) and non-ruptured, non-culprit and hence essentially stable plaques.

So to summarize, we have evidence that:

1. Wrong vessel PCI is very common, perhaps nearly a third of patients (Heitner et al.)
2. Stenting stable lesions in most patients does not reduce MACE (COURAGE, BARI 2D, ISCHEMIA, REVIVED-BCIS2)
3. Stenting all lesions (stable and unstable) in patient with ACS does reduce MACE (COMPLETE, MULTISTARS AMI)

It begs the question of whether the benefit seen in COMPLETE and MULTISTARS AMI could be mediated in part by reducing or eliminating the possibility of missing the culprit?

All the more reason to use intravascular imaging (IVUS or OCT) to identify plaque rupture and confirm the culprit, especially since the use of intravascular imaging was shown to reduce all-cause mortality by 25%.

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

MY Comment, by KEN GRAUER, MD (3/25/2025):

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

I'd suggest as one of the "themes" for Dr. Frick's thought-provoking case from today being — Back to Basics — with not following this fundamental concept leading to the unfortunate occurrences in today's case.

• In a higher-risk patient (as today’s patient was) — If the initial ECG is laden with artifact that prevents accurate interpretation — then immediately repeat the ECG. Instead — no ECG was repeated for the rest of the day.
• Even if no artifact at all would have been present — in a higher-risk patient with ongoing symptoms, ECGs should be repeated every 15-30 minutes until a definitive diagnosis can be made.
• As described above by Dr. Frick: Morphine was given — Troponins were elevated and climbing — yet cardiac cath was delayed (being performed only after a prolonged resuscitation that perhaps could have been avoided had the basics been followed).
• And, perhaps because multi-vessel disease was found on cath with no clear "culprit" being evident — PCI was applied to the wrong artery. As per Dr. Frick — use of intravascular imaging (IVUS or OCT) could have helped to find plaque rupture and identify the "true" culprit. 
• Following through with frequent serial ECGs on this patient might also have helped to identify "tell-tale" ST-T wave changes that would have identified the vascular area(s) contributing to the acute event.

The Initial ECG:

Given how much depends on assessment of the initial ECG — this initial ECG should be immediately repeated if it is not adequately interpretable.

• To illustrate this point in Figure-1 — I've reproduced today's initial tracing in the TOP panel. Baseline artifact is clearly distorting ST-T wave morphology in each of the 6 limb leads. 
• QUESTION: Which of the 3 or 4 QRST complexes in leads II,III,aVF — and in lead aVL — manifest the "true appearance" of ST-T wave morphology in these leads?

In the BOTTOM panel of Figure-1 — I picked the middle QRST complex.

• In today's 60+ year old woman with new-onset severe CP (Chest Pain) — IF the ST-T waves that I show in the BOTTOM panel are the "true" representation of ST-T wave morphology in these 4 leads — Isn't this worrisome? 
• At least in leads II and aVF — I thought the T waves looked disproportionately larger-than-expected (possibly hyperacute) — with what appears to be reciprocal ST depression in lead aVL.
• In this context of new-onset severe CP and potential inferior OMI — I thought the clearly abnormal ST segment flattening in lead V2 and the more subtle (but nevertheless abnormal) ST segment straightening in lead V3 was in support of my concern about an infero-postero OMI.

How do we know IF I picked the "true" ST-T wave morphology?

• Answer: We don't know — because "garbage in = garbage out" — and there is simply too much artifact to tell. Perhaps if this initial ECG had been immediately repeated — and followed up by a 2nd ECG within 15-20 minutes — we might have seen enough "dynamic" ST-T wave change to qualify as an acute event in need of prompt cath (and the VFib episode might never have occurred).

Figure-1: I've labeled the initial ECG in today's case.

When ACS care is very delayed, "STEMI metrics" can be perfect. And how specific is Queen of Hearts?

 Written by Pendell Meyers, sent by anonymous, with additions by Smith

A man in his 40s had acute chest pain and called EMS.

EMS arrived and recorded this ECG:

What do you think?

Here is the PMcardio Queen of Hearts interpretation of the ECG:

STEMI equivalent detected. Inferior and posterior OMI without STEMI criteria.

If you think the Queen of Hearts is so sensitive because it sacrifices specificity, you would be wrong.

See this article, in which using the Queen would decrease false positives substantially:

Artificial Intelligence Driven Prehospital ECG Interpretation for the Reduction of False Positive Emergent Cardiac Catheterization Lab Activations: A Retrospective Cohort Study

There is much more data that shows that the Queen of Hearts is both more sensitive AND more specific (see 3 studies at the bottom)

Case Continued

The ED physician reviewed the prehospital ECG as "No STEMI” at triage and did not order a repeat.

The patient was placed in the waiting room. 

About an hour later another ECG was obtained:

Barely meets STEMI criteria in inferior leads, but obvious inferior and posterior OMI. Even lateral leads V5-V6 have HATW.

The cath lab was now activated for STEMI.

Cath done around 4 hours after first medical contact (symptom onset time uncertain) revealed an RCA occlusion requiring thrombectomy and PCI. 

Smith: How much does such a delay matter?  See this graphic here from JAMA 2005, by legendary authors Gregg Stone and Harvey White. You can read the abstract (or of course the entire article) at the link provided.  Very interesting.  

SUMMARY: From onset of symptoms, a 2 hours delay, from hour 2 to hour 4, results in loss of 60% of the mortality benefit of reperfusion

Gersh BJ, Stone GW, White HD, Holmes DR Jr. Pharmacological facilitation of primary percutaneous coronary intervention for acute myocardial infarction: is the slope of the curve the shape of the future? JAMA [Internet] 2005;293(8):979–86. Available from: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=15728169   

The "STEMI metrics" for this patient reflect a "successful" time to cath within 90 minutes of the STEMI+ ECG. So the final diagnosis is STEMI, and there is no question or feedback about whether the acute coronary occlusion could have been identified sooner. 

If I were the patient, I would want intervention to be 2 hours earlier, wouldn't you?

What would you want? Does the disease only matter when it becomes easy enough to see on the ECG without any dedicated training? Is the disease the ST segment? or is the disease acute coronary occlusion?

In our study below, among patients who actually had STEMI at some point on the ECG, the PMCardio Queen of Hearts AI Model was able to detect OMI a median of 3 hours earlier than STEMI criteria. 

(And, in addition, she was able to detect many OMI that were never STEMI)

Herman, Meyers, Smith, et al. International Evaluation of an Artificial Intelligence-Powered Electrocardiogram Model Detecting Acute Coronary Occlusion Myocardial Infarction. EHJ DH. 2023.

https://doi.org/10.1093/ehjdh/ztad074

McLaren et al.  Missing Occlusions: Quality gaps for ED patients with Occlusion MI.  

More on Specificity of the Queen of Hearts:

1)

Perfect sensitivity and you can see the false positives at the bottom of the slide

2)

Look at the false positives on the right:  

For the cardiologist, none of the FP had a culprit.  

For the Queen, all 4 FPs had a culprit but the troponin was not high enough to qualify for this very strict OMI outcome definition requiring a peak hs troponin I of at least 5000 ng/L.

3)

https://herzmedizin.de/fuer-aerzte-und-fachpersonal/kongresse/dgk-jahrestagung-2025/-vor--kongressprogramm/wissenschaftliches-programm/koronare_herzerkrankungen/v1146.html

Artificial intellingence based detection of acute coronary occlusion compared to STEMI criteria - External validation study in a consecutive all-comer German chest pain unit cohort.

In ROC analysis the AI model had an area under the curve of 0.978 detecting ACO (see C). The model’s sensitivity was 70.2%, its specificity was 98.5%. This resulted in a negative predictive value of 99.2%. STEMI criteria reached an area under the curve of 0.92. The corresponding sensitivity was 28.8%, while the specificity was 93.7%. The negative predictive value of STEMI criteria was 98.1%.

In the subgroup of low risk patients ruled-out by hs-cTnT algorithm (n=2999) the AI model generated false positive results in 0.7% (n=20, specificity: 99.3%) of cases, compared to 5% (n=150, specificity: 95%) using STEMI criteria.

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

MY Comment, by KEN GRAUER, MD (3/23/2025):

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

Although details from today’s case are limited — the “answer” should be obvious within seconds on seeing today’s initial ECG (TOP tracing that I have reproduced and labeled in Figure-1). This man in his 40s with acute CP (Chest Pain), severe enough to prompt him to call EMS — has had an infero-postero OMI that has-to-be assumed acute until proven otherwise.

• Learning more precisely when this patient’s CP began — and how severe his CP was at the time ECG #1 was obtained (as well as learning if this patient has a previous history of heart disease and what his “baseline” ECG looks like) — would help us to better determine the likely onset of his OMI.
• That said — there are a number of findings in this patient’s initial ECG that look recent, if not acute.

Why this Patient should NOT be sent to the Waiting Room:

After hearing the history and seeing the initial ECG in Figure-1 — today’s patient was sent to the Waiting Room. We are lucky this patient did not have a cardiac arrest while in the Waiting Room. Instead (as per Dr. Meyers) — worrisome ECG findings were evident on the initial ECG.

• My “eye” was immediately drawn to leads V2 and V3 (within the RED rectangle in ECG #1). In a patient with new CP — early transition (with a predominant R wave already in lead V2) — in association with loss of the normal slight, gently upsloping ST segment in leads V2,V3 (replaced by ST segment flattening — as highlighted by the RED arrows) is diagnostic of acute posterior OMI.
• Infarction Q waves are seen in each of the inferior leads (most notably in lead III). Considering the modest inferior lead QRS amplitudes — acuity is suggested by disproportionately "bulky" T waves in leads II and aVF (BLUE arrows in these leads). Even without reciprocal ST depression in lead aVL — in this patient with new CP, these findings are diagnostic of acute inferior OMI.
• KEY Point: If there was doubt about whether the above ECG findings are diagnostic — then ECG #1 should have been repeated within 15-20 minutes, rather than sending the patient to the Waiting Room, and not repeating the ECG for an hour. 

Figure-1: Comparison between the 2 ECGs in today's case. (To improve visualization — I've digitized the original ECG using PMcardio).

The Repeat ECG:

As per Dr. Meyers — lead-to-lead comparison of ECG #2 with ECG #1 leaves no doubt about the diagnosis of an acutely evolving infero-postero STEMI.

• Learning Point: The BEST way to improve our ability at recognizing less obvious but important subtle findings early on, is to go back to the initial ECG and compare lead-by-lead how the initial ECG evolved. That the disproportionately large T waves in leads II and aVF of the initial tracing were hyperacute is now proven by what the ST-T waves in these leads have become. Even the subtle ST segment straightening in lead III of the initial ECG was a harbinger of the marked ST elevation now seen in lead III of ECG #2.
• Similarly — the subtle-but-real ST segment straightening highlighted in lead V2 of ECG #1 — has now evolved to a more acute-appearing downsloping ST segment in ECG #2.
• Sombering Reality: Just as we saw in the recent March 17, 2025 post in Dr. Smith's ECG Blog — "quality control" of today's case will view physician recogition of STEMI criteria in ECG #2 as a "correct" interpretation — whereas STEMI criteria should never have developed because the cath lab should have been immediately activated as soon as the physician saw ECG #1.