The diagnosis is NSTEMI, a worthless diagnosis which harms patients.

 Written by Magnus Nossen, with many comments by Smith

A 50-something male with a medical hx of type I diabetes and hypertension presented to the ED with complaints of chest pain. The ECG below is recorded. 

• What are your concerns if any, and how would you manage this patient?
  

6 x 2 format
ECG #1 recorded on admission

I sent this to Dr. Smith with no information, and he responded immediately with "SAF, subtle but aVL is unmistakable."

"SAF" means "South Africa Flag" sign.  This is a sign of First Diagonal Occlusion (OMI).  See below.  You actually can't see the SAF in 6x2 format. You will see it below after transformation to the 3x4 format.

First of all this patient has a very high pre-test probability of having a myocardial infarction as he has type I diabetes, hypertension, and is now presenting with acute onset chest pain. Despite not showing any significant ST segment elevation — this ECG is diagnostic of OMI. Expert interpreters who are trained in recognizing OMI will have no problem stating with a high degree of confidence that this ECG is consistent with occlusion myocardial infarction. As with many cases of OMI, the ST segment changes are subtle. Rather than relying on absolute ST segment change and millimeter criteria — it is the sum of all findings that make the ECG diagnostic. 

Because the ECG does not have "diagnostic" ST Elevation, this patient was admitted with a diagnosis of NSTEMI, but only after an admission troponin I returned elevated at 102ng/L (ref value <34ng/L). The repeat troponin I after 3 hours returned at 429ng/L.  The patient was given morphine for pain, which decreased the pain, but did not resolve it.  Morphine is a terrible mistake unless you are already committed to the cath lab.

"NSTEMI" implies that the patient does not need emergent intervention, even when it is an OMI which does need emergent intervention (see long discussion below).

Actually, however, NSTEMI with persistent pain is a universally agreed upon, and guideline recommended, indication for intervention, but it is rarely followed (only 6% of the time in the only study).

_________

Morphine abstract:

Bracey, A.  Meyers HP.  Smith SW.  Wei L. Singer DD.  Singer A.  Association between opioid analgesia and delays to cardiac catheterization of patients with occlusion Myocardial Infarctions. Academic Emergency Medicine 27(S1): S220; May 2020.  Abstract 556.

Main result: STEMI(-) OMI Patients

65 (23.9%) patients were found to have STEMI(-) occlusion myocardial infarction (OMI) at the time of cardiac catheterization. The 45 patients with STEMI(-) OMI without pre-cath opioids had a door-to-balloon time of 75 minutes, vs. 684 minutes for the 25 STEMI(-) OMI with pre-cath opioids.

Let us have a closer look at the admission ECG. The ECG shows a very subtle South African Flag (SAF) sign. The SAF sign is associated with high lateral infarcts, and is often due to occlusion of the first diagonal branch of the LAD. Below in Figure 1, the South African Flag has been overlayed on the initial ECG in today's case.

This ECG pattern is characterized by ST segment elevation in leads I, aVL and V2, with ST depression in lead III. In today’s ECG — the leads that usually show STE in the SAF sign, manifest clear ischemic change but without frank ST elevation. Lead I in our case has significant baseline wander — but several of the QRS complexes in that lead clearly show hyperacute appearing T waves. Lead aVL is very pathologic and highly concerning if not diagnostic for OMI by itself. It shows minimal upward concave ST segment elevation and a (when compared to the preceeding ECG) bulky, hyperacute T waves. Lead III has slight ST depression and an inverted T wave which is reciprocal to the high lateral hyperacute T wave and ST change in aVL. Lead V2 shows a symmetric ischemic looking T wave — and the remaining precordial leads have ischemic looking minimal ST depression. 

Here I have re-formatted the ECG from 6x2 to 3x4, which many find easier to interpret.  

It is also the only format which shows the South Africa Flag!!

Figure-1: ECG #1 overlayed with the South African Flag

The SAF sign is considered an uncommon but important ECG finding, as high lateral occlusions are relatively uncommon and this ECG pattern can easily be missed due to its non-contiguous ST-segment changes.

These findings were not appreciated by medical providers on this case. The following morning, the below ECG was recorded after high sensitivity troponin I returned at just above 14.000 ng/L.   As we can see there is loss of R waves in leads V2 and aVL. Leads I, aVL and V2 show newly developed pathologic Q-waves. 

ECG #2 recorded 13 hours after admission

The patient eventually had an angiogram done. 

This was done close to 24 hours after symptoms started, and the prognostic value of doing PCI for patients with continuous occlusion for nearly 24 hours is limited. At cath, an occlusion of the 1st diagonal branch of the LAD was opened and stented. 

The PMCardio Queen of hearts ECG AI model was not used in the management of this patient. Had the AI model been implemented in this patient's care — it would have identified the first ECG as an OMI, resulting in a significantly reduced symptom-to-balloon time. 

The Below is the Queen of hearts interpretation and explainability feature. As you can appreciate, the explainability feature highlights the ECG changes in a SAF pattern. 

The Queen of Hearts applies the label "STEMI/STEMI equivalent" to anything she recognizes as an acute coronary occlusion (OMI).  

New PMcardio for Individuals App 3.0 now includes the latest Queen of Hearts model and AI explainability (blue heatmaps)! Download now for iOS or Android.

Learning Points:

• Sudden onset chest pain in a patient with high pre-test probability of ACS is myocardial ischemia until proven otherwise. 
• Acute coronary occlusion requires immediate angiography and intervention.
• ECG changes in OMI can be subtle. In such case, especially when combined with an unusual ECG lead distribution — acute ischemic changes can be difficult to appreciate.

Smith's most important learning points:

1.  "NSTEMI" is a worthless diagnosis which encompasses both patients who need (OMI), and those who do not need (Non-OMI, or NOMI), emergent intervention.

2. Never give morphine unless you are committed to the cath lab

3. Contrary to widespread belief, patients with "NSTEMI" do benefit from emergent intervention.  See trials summarized below.  It is a myth that they do not, a myth propagated because many trials excluded patients with persistent symptoms and because "early" was not really early, but rather 5-16 hours after trial enrollment, which is generally 7-18 hours after ED presentation.  This is much too late for any benefit.

4. "NSTEMI" is usually identified by a "Nondiagnostic" ECG and elevated troponins.  When the NSTEMI is OMI (as opposed to NOMI), most such ECGs are recognizable as OMI by an expert or by the Queen of Hearts.  When it is not recognized, the diagnosis made by troponins is too late to make a BIG difference, and so many of these trials are negative.

We know that the benefit of reperfusion drops precipitously with just a 2 hour delay:

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

NSTEMI Trials: "Early" vs. Late PCI in NSTEMI

Here is a table summarizing all trials: 

https://docs.google.com/document/d/105SJvgOeoYnpSWDxA8_03ioNKUepwhaRs892rwRw5Ec/edit?usp=sharing

Here is a quick summary:

Individual Trials

TIMACS (Timing of Intervention in ACS) 

Mehta, et al., New England Journal

2009

•N= 3,000 NonSTEMI randomized to early or routine PCI

•Results: No difference in death/MI/stroke

•GRACE score > 140: better outcome with early PCI

•“Early” intervention was not early

–Median 16 vs. 52 hours

•Methods: no mention of refractory ischemia

•Personal communication with Dr. Mehta

–“I doubt investigators would have enrolled patients with ongoing chest pain in this Non-STE-ACS trial.”

VERDICT Trial (Early Versus Standard Care Invasive Examination and Treatment of Patients with Non-ST-Segment Elevation Acute Coronary Syndrome)

Kofoed et al, Circulation 2018

N=2147 NSTEMI pts (exclusions below), randomized to "very early" (4.7 hrs) vs. "standard" (62 hrs) angiography

88% vs. 84% in each group underwent coronary revascularization

median follow-up time 4.3 years

Primary endpoint: combination of all-cause death, nonfatal recurrent MI, admission for refractory myocardial ischemia, admission for heart failure

- All pts: 27.5% in very early group, 29.5% standard care, no difference

 - Pts with GRACE>140:

GRACE score > 140 confirmed in VERDICT trial

4.7 vs. 62 hours, persistent symptoms excluded. 

Kofoed KF, Kelbæk H, Hansen PR, Torp-Pedersen C. Correction to: Early Versus Standard Care Invasive Examination and Treatment of Patients With Non-ST-Segment Elevation Acute Coronary Syndrome: VERDICT Randomized Controlled Trial. Circulation [Internet] 2018;138(24):e750. Available from: http://dx.doi.org/10.1161/CIR.0000000000000640

Exclusion criteria were pregnancy, patient inability to understand trial information, an indication for acute ICA (very high-risk NSTE-ACS,12 including ongoing ischemia despite intravenous nitroglycerin infusion, hemodynamic or electric instability, acute heart failure, mechanical complication, or cardiac arrest), expected survival 

RIDDLE-NSTEMI

Milosevic A, Vasiljevic-Pokrjcic Z, Milasinovic D, et al. Immediate Versus Delayed Invasive Intervention for Non-STEMI Patients. Journal of the American College of Cardiology Cardiovasc Interv 2016;9(6):541–9.

ABOARD* trial (n=360), Montalescot, JAMA 302:947; 2009

–       1 vs. 20 hours, no diff., LMWH mostly, No sophisticated ECG analysis

–       Refractory symptoms excluded

TIMACS,* Mehta, NEJM 360:2165; 2009, 3000 patients

–       No diff except for patients with in-hospital GRACE score > 140 (Early is better)

–       Early was not very early (16 vs. 50 hours)

–       Excluded refractory ischemia (personal communication, Mehta)

ISAR-COOL,* JAMA 290:1593, 2003; n = 410

–       Better early (but 2.4 hours vs. 86 hours); included tirofiban and clopidogrel 600

–       Death or Large MI: 11.6% vs. 5.9%;

–       Before PCI: 3 death and 10 large MI vs. 0 deaths and 1 small MI; After PCI: 11 MI in both groups

LIPSIA-NSTEMI, Thiele, Eur Ht. Journal 2011, n = 400

–       1.1 vs 18.6 hours.  Death or MI: 21% (delayed) vs. 16% (Immediate), p = 0.17, low power

–       Excluded refractory ischemia

ELISA PCI, van’t Hof, Eur Ht J 24:1401; 2003.  N= 220. Delayed always got tirofiban. All LMWH.

–       6 vs. 50 hours.  Patent vessel: 66% (late) vs. 82% (p = 0.05)

Sisca trial: Reuter P.  Int J Cardiol 182:414; 2015 Did not exclude persistent symptoms

–       Early invasive far better. Median 2.8 hrs (1.4-4.7)

Verdict trial (2018): Early 4.7 hrs median. Ongoing symptoms were an exclusion

Optima: within 24 hours vs. 24-72 hours.  Early better.  Multiple studies show 24 hours is better than 72.

Meta-analysis.  Katritsis Eur Ht. J 32:32; 2011 (excludes LIPSIA-NSTEMI)

–       In non-urgent cases, early PCI reduces recurrent ischemia and shortens length of stay

Summary: Rapid/Urgent Cath for high risk, unstable, or refractory ischemia.

  

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

MY Comment, by KEN GRAUER, MD (5/21/2025):

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

Today's post by Dr. Nossen is indeed subtle. Making interpretation of today's initial ECG even more challenging — is the "technical misadventure" of artifact plus baseline wander that is embedded in this tracing (For clarity — I've reproduced this initial ECG in Figure-1.)

• My comment below is focused on honing the ability of emergency providers to sort out the camouflage effect of these technical misadventures — that otherwise might serve to mask and conceal subtle-but-essential ECG findings in ECG #1.

Which Beats Should We Look At?

There are numerous different ECG formats — each with its own pros and cons.

• The advantage of the recording format seen in Figure-1 — is that we see more beats in each lead (ie, up to 5 seconds in each of the 12 leads — instead of the usual 2.5 second segments when the ECG format records 4 groups of 3 simultaneous leads).
• My "eye" is used to the 4 groups of 3 leads format, in which a simultaneously-recorded 10 second long lead II rhythm strip appears under the 12-lead. I favor this format because I find it easier to quickly assess comparative findings in each of the 12 leads.
• BUT — Aren't there a number of limb leads in Figure-1 that manifest a distracting amount of artifact and baseline wander?

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

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

The KEY ECG Findings:

As per Dr. Nossen's discussion above — today's patient is a 50-something year old man, who presents to the ED with new CP (Chest Pain). 

• For me — the KEY finding that caught "my eye" (as seen in Figure-2) — are ST segments for the 4 beats within the RED rectangle in lead aVL (beats #5,6,7,8). Confirmation that the ST segment straightening for these 4 beats truly represents a hyperacute finding — is forthcoming from the reciprocal ST depression in the inferior leads (BLUE arrows in leads II,III,aVF).
• Acknowledgment: I would have had a different impression regarding ST-T wave assessment in lead aVL if the ECG I was shown had only recorded 2.5 second segments, using the first 4 beats in lead aVL.

Chest lead findings in Figure-2 are even more subtle.

• As per Dr. Nossen — there is no ST elevation in lead V2. But, given the clearly abnormal (hyperacute) ST-T wave in lead aVL — the T wave in lead V2 looks disproportionately tall, and is clearly "fatter"-at-its-peak than expected given modest size of the QRS in this lead.
• There also is no frank depression in the remaining chest leads. Instead, there is ST segment straightening in neighboring leads V3,V4 (BLUE arrows), and to a lesser exent in leads V5,V6.

Figure-2: I've labeled KEY ECG findings in today's initial ECG.

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

IF the Initial ECG was as shown in Figure-3:

To better illustrate the potential distracting effect of today's technical misadventures — Look at Figure-3 — in which I reduced the length of each lead recording (showing simultaneously-recorded beats #1,2,3,4 in the limb leads — and beats #9,10,11 in the chest leads).

If the 12-lead recording of today's initial ECG only used the complexes shown in Figure-3: 

• I would clearly have been distracted by the artifact and baseline wander in the limb leads. There would have been too much variability of the ST-T wave in lead aVL for me to appreciate any clear abnormality.
• ST depression is present in the inferior leads — but there once again is so much variability from artifact and baseline wander — that I could not with any confidence suspect an acute OMI.
• I also would have thought that lead V2 was inconclusive — with obvious artifact distortion of the 1st ST-T wave in this lead — and an almost normal-appearing ST-T wave in the 3rd ST-T wave.
• I would not have perceived leads V3,V4,V5,V6 as showing anything conclusive.

Figure-3: IF today's initial ECG looked like this ...

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

IF the Initial ECG was as shown in Figure-4:

On the other hand — I'd arrive at an entirely different assessment if today's initial ECG looked like Figure-4 (in which each lead recording was shortened — showing simultaneously-recorded beats #5,6,7,8 in the limb leads — and beats #12,13,14,15 in the chest leads).

• In today's patient who presents with new CP — my "eye" would immediately be captured in Figure-4 by the clearly abnormal (hyperacute) ST-T wave that is consistently seen for the 4 beats in lead aVL.
• In this context — I would also interpret the small, but clearly broader-than-expected T wave in lead I as a hyperacute change.
• Confirmation of these hyperacute high-lateral lead changes — would clearly be forthcoming from consistent ST-T wave depression in each of the inferior leads.
• Given the above ST-T wave abnormalities in the 5 limb leads — I'd interpret the "bulkier"-than-expected T wave in lead V2 as a hyperacute change — supported by clearly abnormal ST segment flattening in leads V3,V4 (and to a lesser extent in V5,V6).
• NOTE: There is essentially no baseline wander in Figure-4. Artifact in this tracing is minimal — and does not detract from accurate interpretation of this tracing!

Figure-4: IF today's initial ECG looked like this ...

 ===============================

Why I Prefer the 4 Groups of 3 Leads Format:

In Figure-5 — I've "reconstructed" today's initial ECG by selecting 4 consecutive beats in each of the 12 leads from Figure-1, in which there is essentially no artifact or baseline wander. Isn't it much easier to interpret this ECG in Figure-5?

• We periodically review examples of the South African Flag Sign — in which the pattern of: i) ST elevation in leads I,aVL,V2; + ii) Reciprocal ST depression in lead III; — but — iii) No ST elevation in any chest leads except for lead V2 — indicates acute occlusion of either the 1st or 2nd Diagonal branch of the LAD (See My Comment in the January 18, 2025 post, among many others).
• Although today's initial ECG lacks the ST elevation needed to fully satisfy criteria for the South African Flag Sign — the hyperacute T waves in leads I,aVL,V2 — in association with ST depression in lead III and ST flattening in leads V3-thru-V6 strongly suggests a 1st or 2nd Diagonal as the "culprit" artery. This was confirmed on cardiac cath in today's case.

Figure-5: Wouldn't interpretation be much easier if this was the initial ECG?

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

Moral of Today's Story:

• Be aware of "technical misadventures" such as artifact and baseline wander — that may render subtle-but-essential ECG findings almost impossible to recognize.
• In my experience — one of the most common oversights by even experienced clinicians is the failure to immediately repeat the ECG when tremor, excessive patient movement, labored breathing, loose electrode leads, or other artifact-producing action hinders accurate interpretation in a patient with new CP. If clinical decision-making is dependent on determining if an acute cardiac event is ongoing — Try your best to mitigate artifact-producing actions — and immediately repeat the ECG!
• 
  
• P.S.: Ready review of numerous examples of Technical "Misadventures" is available by clicking on the 5th item in the lower Menu at the TOP of every page in Dr. Smith's ECG Blog (CLICK HERE).