Friday, November 3, 2017

12 Example Cases of Use of 3- and 4-variable formulas, plus Simplified Formula, to differentiate normal STE from subtle LAD occlusion

Here are 12 Example Cases using the 3- and 4-variable formulas, which help to distinguish normal ST elevation in leads V2-V4 from that of subtle LAD occlusion.

WARNING: the formula is not perfect.  Beware of using it to reverse your prior opinion that the ECG represents LAD occlusion.  I recommend using it when you are worried that an ECG with apparent normal ST Elevation might be LAD occlusion.  Sensitivity is not perfect.

New warning: when using QTc, do NOT correct if the heart rate is below 60 beats per minute.

Instructions for using free subtleSTEMI iPhone app for the 3-variable formula.  4-variable version still to come.

Here is a video lecture of subtle LAD occlusion: One hour lecture on Subtle ECG Findings of Coronary Occlusion

The 3-variable formula comes from this paper:
Smith SW et al.  Electrocardiographic Differentiation of Early Repolarization FromSubtle Anterior ST-Segment Elevation Myocardial Infarction.  Annals of Emergency Medicine 2012;60:45-56.

The 4-variable formula comes from this subsequent paper:

It was externally validated in this study:
A tale of two formulas: Differentiation of subtle anterior MI from benign ST segment elevation

It is critical to use it only when the differential is subtle LAD occlusion vs. early repol. Thus, there must be ST Elevation of at least 1 mm in any of leads V2-V4. If there is LVH, it may not apply. If there are features that make LAD occlusion obvious (inferior or anterior ST depression, convexity, terminal QRS distortion, Q-waves, or ST Elevation of 5 mm or more), then the formula MAY NOT apply. These kinds of cases were excluded from the study as obvious anterior STEMI.
--QTc is the computer measurement.  Do not correct for heart rate if it is under 60
--RAV4 = R-wave amplitude, in mm, in lead V4.
--ST elevation (STE) is measured at 60 milliseconds after the J-point, relative to the PR segment, in millimeters.

Formula: (1.196 x STE at 60 ms after the J-point in V3 in mm) + (0.059 x computerized QTc) - (0.326 x R-wave Amplitude in V4 in mm).
A value greater than 23.4 is quite sensitive and specific for LAD occlusion.
A value greater than 22.0 is more sensitive, has fewer false negatives, but specificity drops to 80%.

4-variable formula (slightly more accurate than 3-variable):

Uses the same measurements, but adds the total QRS amplitude (both R- and S-waves) in lead V2:

Formula: (1.062 x STE at 60 ms after the J-point in V3 in mm) + (0.052 x computerized QTc) - (0.151 x QRSV2) - (0.268 x R-wave Amplitude in V4 in mm).
A value greater than 18.2 is quite sensitive and specific for LAD occlusion.

The 4-variable formula is based on this paper:  

Driver, BE et al.  A new 4-variable formula to differentiate normal variant ST segment elevation in V2-V4 (early repolarization) from subtle left anterior descending coronary occlusion - Adding QRS amplitude of V2 improves the model.  J Electrocardiology 50(5):561-569; September/October 2017.

Go here to find excel applets. you enter the

The closer the formula value is to the cutpoint of 18.2, the more accurate. 

See this post for more precise interpretation of the results of the 4-variable formula, with graphs.  

In summary: 
At a cutpoint of 17.0, it is 97% sensitive.  
At a cutpoint of 19.0, it is 97% specific.

Finally, there is a simplified formula without QT correction
Aslanger.  Am J Cardiol 2018; 122(8):1303-1309.

(RAV4 in mm + QRSV2 in mm) - [(QT in mm) + STE60V3 in mm)

A value less than 12 corresponds to LAD occlusion; a value greater than 12 to normal variant STE

Sens, Spec, Acc = 86.9%, 92.3%, 90.1%

Case 1

Acute anterior STEMI from LAD occlusion, or Benign Early Repolarization (BER)?

This is the initial ED ECG of a 46 year old male with chest pain:
The QTc was 420
ST Elevation at 60 ms after the J-point in lead V3 = 2.5 mm
R-wave amplitude in V4 = 2
QRS V2 = 19 mm
3-variable formula = 27.11 (above 23.4, consistent with LAD occlusion)
4-variable formula = 21.09 (above 18.2, consistent with LAD occlusion)
This patient was misdiagnosed with normal variant STE.  He was admitted for a "rule out MI." He ruled in for acute MI with a peak troponin I of 130 ng/mL.

Next day angiogram showed a 100% proximal LAD occlusion.  He had a convalescent anterior wall motion abnormality.

Case 2

Acute MI from LAD occlusion, or early repolarization?

The QTc was 455
ST Elevation at 60 ms after the J-point in lead V3 = 3.0 mm
R-wave amplitude in V4 = 17
QRS V2 = 16 mm
3-variable formula = 24.89 (above 23.4, consistent with LAD occlusion)
4-variable formula = 19.87 (above 18.2, consistent with LAD occlusion)

This was not recognized, repeat ECG at 69 minutes showed new Q-waves in V2-V4 that were not appreciated, and patient waited for a prolonged period in the emergency department before STEMI was diagnosed.

100% LAD occlusion.

Case 3

I was reading a stack of ECGs yesterday, and saw this one, with no clinical information....

The QTc was 400
ST Elevation at 60 ms after the J-point in lead V3 = 1.5 mm
R-wave amplitude in V4 = 4.0
QRS V2 = 22 mm

3-variable formula = 24.1 (above 23.4, consistent with LAD occlusion)
4-variable formula = 18.0 (below 18.2, consistent with Normal STE)
The 3-variable formula is positive, but because of a 22 mm QRS in V2, the 4-variable formula is barely negative.  Whenever the value comes close to the cutoff, it is very hazardous to call it negative.  Appropriately, the physicians repeated the ECG 20 minutes later and it was diagnostic of anterior STEMI.

100% LAD occlusion.

Case 4

Transient STEMI, serial ECGs prehospital to hospital, all troponins negative (less than 0.04 ng/ml)

A 45 year old male called 911 for chest pain:
The QTc was 400
ST Elevation at 60 ms after the J-point in lead V3 = 3.5 mm
R-wave amplitude in V4 = 13
QRS V2 = 18 mm

3-variable formula = 23.54 (above 23.4, consistent with LAD occlusion)
4-variable formula = 18.315 (above 18.2, consistent with LAD occlusion)

It was not diagnosed. Fortunately, it spontaneously reperfused and subsequent ECGs show resolution of ST Elevation.  All troponins were below the level of detection.

Angiogram showed a critical LAD thrombotic stenosis.  He underwent CABG.

Case 5. False Negative for both formulas.

Male in early 40's with 1.5 hours of chest pain

This was sent to me for my interpretation without clinical data:
This was my response: 
"Tough one.  Probably is acute LAD occlusion, needs serial ECGs/echo."


The QTc was 410
ST Elevation at 60 ms after the J-point in lead V3 = 2.5 mm
R-wave amplitude in V4 = 14
QRS V2 = 16 mm
3-variable formula = 22.6 (below 23.4, consistent with normal STE)
4-variable formula = 17.8 (below 18.2, consistent with normal STE)
At a cutoff of 23.4 (my typically recommended cutoff), the sensitivity for LAD occlusion among subtle cases is 86% (much higher if all LAD occlusions are used as the denominator), with a specificity of 91%.

At a cutoff of 22.0, the sensitivity was 96% but with decreased specificity of 81%.  I get worried about any value greater than 22.

To me, the T-waves looked too fat to be normal.  And even though the formula was not greater than 23.4, at 22.6 or 23.2, it is close enough to be worried.

At 7 hours after presentation, after several positive troponins, the patient went for an angiogram and had a 100% LAD occlusion.  Peak Troponin T was very high at 2.11 ng/mL.

Case 6.

A 75 year old woman who complains of "Reflux."

Incredible Case Demonstrating the Value of Frequent Serial ECGs

The QTc was 400
ST Elevation at 60 ms after the J-point in lead V3 = 2.5 mm
R-wave amplitude in V4 = 4.5
QRS V2 = 6 mm
3-variable formula = 25.12 (above 23.4, consistent with LAD occlusion)
4-variable formula = 21.17 (above 18.2, consistent with LAD occlusion)
This prompted serial ECGs:
These dynamic changes confirmed ACS

There was a large obtuse marginal 100% occlusion (surprisingly not LAD!  But, in reality, the formula only tells you that the wall under leads V2-V4 is involved; it cannot actually specify the artery involved).

Case 7.

A 60 year old male had resolving chest pain
There is minimal ST elevation and there are subtle T-wave inversions in V2-V5, highly suggestive of Wellens' syndrome.  
Is the ST Elevation normal? Very low R-wave amplitude suggests NOT.  See formula here:

The QTc was 380
ST Elevation at 60 ms after the J-point in lead V3 = 1.5 mm
R-wave amplitude in V4 = 2.5
QRS V2 = 15.5 mm
3-variable formula = 23.4 (equals cutoff of 23.4, consistent with LAD occlusion)

4-variable formula = 18.34 (above 18.2, consistent with LAD occlusion)
So this, along with the resolving chest pain, appears to be a reperfusing LAD occlusion.

In Wellens' syndrome, the artery is open.  T-wave inversion is indicative of reperfusion of the infarct-related vessel.

BP was elevated, pulse lowered to 45.  He received ASA, Plavix 600mg, Heparin, and a Nitro drip.

Another ECG 15 minutes later, pain free, had more T-wave inversion.

Then 15 minutes after that, the pain recurred and the T-waves pseudonormalized.  The ST elevation remained subtle, but the formula value was even higher.

The patient went to cath and had a distal LAD 99% stenosis with thrombus and TIMI-2 flow.

Initial troponin I returned at 1.5 ng/ml.  Peak was 8.1 ng/ml.

Case 8.

Formula positive for LAD occlusion. But echo shows no wall motion abnormality! What is it?

A 35 year old male complained of chest pain:
The QTc was 405
ST Elevation at 60 ms after the J-point in lead V3 = 4.0 mm
R-wave amplitude in V4 = 12.5
QRS V2 = 12.5 mm

3-variable formula = 23.95 (above 23.4, consistent with LAD occlusion)
4-variable formula = 19.53 (above 18.2, consistent with LAD occlusion)

This was sent to me without any clinical information, asking for my opinion.  I was worried about LAD occlusion. The clinicians thought it was normal, but due to elevated troponins, thought it was myocarditis.  There was no wall motion abnormality.  After many hours, the decided that it was appropriate to do an angiogram and they found a distal LAD occlusion which was opened and stented.

100% distal LAD occlusion.

Case 9.

Patient presentation is important

This was a 60-something with acute chest pain:
There is sinus bradycardia at a rate of 44. 

There is 2 mm of ST elevation in V2 and V3 (at the J-point, relative to the PQ junction), which is below the "criteria" for men under 40 (2.5 mm). 

There is upward concavity
There is no ST depression
There are no Q-waves
There is no terminal QRS distortion

So early repolarization is a possibility.

The cardiologists were not impressed by the ECG.

However, there is very suspicious STE in III and aVF, with ST depression in aVL.    To me, this makes the ECG nearly diagnostic of ischemia, though if it is LAD occlusion, there should be ST depression in III and aVL, so it is a bit confusing.  Unless you consider a wraparound LAD to the inferior wall. 

There is also an upright T-wave in V1, larger than V6.  This is a very soft sign of LAD occlusion, and did not add any diagnostic improvement over our formula model for diagnosing LAD occlusion vs. early repol.  In that study, this finding was present in 39% of LAD occlusion and 15% of early repol.

The measurements are:

RA V4 = 6.5
STE60V3 = 3.0
QRSV2 = 12.5
QTc = 362 ms

The formula values are thus:

3-variable: 22.83 (less than 23.4; the most accurate, but not most sensitive, cutoff).  I still worry when the value is above 22.0 (still misses 4% of LAD occlusion when below 22.0)

4-variable 18.38 (greater than 18.2; the most accurate, but not most sensitive, cutoff). So this is strongly indicative of LAD occluson.

Computerized QTc = 362 ms (this is VERY short for LAD occlusion, but is so short because of correction for slow heart rate; the measured QT was 413 ms.)  This is the main reason the formula value is not very high.

Thus, the ECG is early repolarization by the old 3-variable formula and is LAD occlusion by the new one.

Even though the ST Elevation is not diagnostic, and T-waves of this size can be seen in early repolarization, but low QRS and R-wave amplitude is not. 

The inferior ST findings also make this ECG diagnostic.

In case you were wondering about the T-waves and bradycardia, the K was normal.

Why bradycardia?  Think about that inferior ST elevation.  Maybe there is also inferior MI from wraparound LAD with associated sinus bradycardia.


The cath lab was activated by the emergency physician. The patient had a 100% LAD occlusion and a peak troponin I of 51 ng/mL.  

Case 10.

A 50-something with chest pain and minimal precordial ST elevation

This ECG was recorded at 1350, with pain decreasing after nitroglycerine:
There is 1 mm of ST elevation at the J-point in both V2 and V3 (within normal limits). 
Computer interpretation is normal
Cardiologist overread is normal
What do you think?

More description: There is also poor R wave progression, with small R waves in V4. The T waves are slightly broad and large, but probably could not be called hyperacute. There is minimal STD in aVF.

One of our interns had texted this ECG-1 to me, with the message:

"3 hours of chest pain, QTc = 415 ms, 3 variable formula is 25.3.  What do you think?"

[The 3-variable formula for differentiating normal variant ST elevation from the ST elevation of subtle LAD occlusion can be accessed by clicking on the link at the top of the page and entering the values into the online excel applet.  Values are: 1. ST elevation at 60 ms after the J-point in lead V3.  2. R-wave amplitude in V4. 3. computerized QTc.  See also the free iPhone app "SubtleSTEMI".  See also MDcalc.]

A value greater than 23.4 is very worrisome for LAD occlusion.

[The 4-variable formula adds the entire QRS amplitude in lead V2 and is more accurate than the 3-variable formula.  It is: (1.062 x STE at 60 ms after the J-point in V3 in mm) + (0.052 x computerized QTc) - (0.151 x QRSV2) - (0.268 x R-wave Amplitude in V4 in mm).  It can also be accessed at the top of the page, with value entry into the excel applet.]

The publication of the formula can be found here: A new 4-variable formula

A value greater than 18.2 is quite sensitive and specific for LAD occlusion. 

The calculated value was 19.2 (highly suggestive of occlusion)

I suggested serial ECGs, which were done as the patient had diminishing pain after nitroglycerine:

Here are the successive ECGs, V1-V3
The first image is of an earlier ECG, of which the providers were unaware.

                   1300               1350, pain decreasing               1430                  1700, pain free       1800, still pain free
This shows that the T-waves (which never were quite hyperacute), are deflating and may have been hyperacute prior to arrival.


The third troponin I, drawn 4.5 hours after presentation, returned at 4.2 ng/mL.

The patient went for angiogram and had an 80% mid-LAD thrombotic stenosis and proximal LAD disease, as well as a 90% diagonal lesion.  He went for Coronary bypass (CABG). 

This outcome is perfectly consistent with all the ECGs.

Learning points:

1. Pay attention to diminishing T-wave amplitude during diminishing pain.

2. The formulas are very accurate.  I have always thought that I can do better than my formulas, but now I'm in doubt.

3.  Some patients have near zero ST elevation at baseline. Any ST elevation in these patients is abnormal.  In such patients, LAD occlusion may result in very subtle ST elevation.

Case 11.

A Male in his 40's with Decreasing Chest pain - what do you think?

A male in his 40s presented with decreasing chest pain.  Here was his first ECG:
There is sinus rhythm. 
The QTc is 379 ms. 
There is ST elevation in V2-V4 that does not meet STEMI "criteria" of at least 1 mm in 2 or more leads (except V2 and V3, which require 2.0 mm of more for men over age 40).

So it looks like early repolarization. 
The clinicians used the subtleSTEMI formula.  I am not sure exactly what numbers they used, but they told me the value they arrived at was 21.2.

Let's examine that:
The computerized QTc was 379 ms.
The R-wave amplitude in lead V4 is from 11-14 mm, depending on the complex used.
Here is a magnification of V3 in order to measure ST elevation at 60 ms after the J-point in lead V3:

The black arrow shows the J-point.  The red arrow is at 60 ms (1.5 small boxes) after the J-point.  The lower edge of the upper green line is where one should measure from.  The upper edge of the lower green line is at the PQ junction.  The distance between these two is 4 mm.  Some might say 3.5 mm

If we put these values into the formula, using QTc of 379, we get 6 different values depending on the measurements:

                                   RAV4 = 11 (min)                     RAV4 = 12.5 (avg)                  RAV4 = 14 (max)
STE = 3.5 mm                   22.96                                       22.5                                             22.0

STE = 4.0 mm                   23.56                                       23.1                                             22.6

Only one of these 6 values is very specific for LAD occlusion (23.56, greater than 23.4)

All the rest are greater than 22.0, above which one should definitely be worried and get serial ECGs.

QRS V2 = 15.5 mm; the 4-variable formula is at the very least 17.6 and at most 18.9. With the most accurate cutpoint at 18.2, this is also ambiguous.

So a second ECG was recorded 12 minutes later:
QTc is now 383 ms.  STE 60 V3 = 2.5 mm.  RAV4 = 11-12 mm. 
3-variable formula = 21.7
4-variable formula = 17.0
Thus, values of both have fallen
Formula value is now slightly lower. 

These two ECGS are significantly different, but it was not noticed the treating physicians.

Look at the ECGs side by side.  And remember the pain is waning.
The first is on the left, the follow up is on the right.
The T-wave amplitude in V3 is 10 mm on the earlier and only 7 mm on the later one.
This changed the ST elevation at 60 ms after the J-point from 4.0 to 2.5.

This makes it almost certain that the ST elevation on the first one is due to ischemia.

Although the clinicians were uncertain and obtained a very low formula value for both, they were worried about the patients symptoms and appropriately activated the cath lab.

The angiogram showed a 99% thrombotic occlusion with TIMI-II flow (enough coronary flow to prevent outright ST elevation).

It is probable that the artery was fully occluded at the time of maximal chest pain.

Learning Points:
1. Hyperacute T-waves diminish in size as the artery reperfuses
2. The formula is more likely to be falsely negative when there is a reperfusing artery.
3. A value less than 23.4 but still greater than 22.0 may still be due to LAD occlusion or near-occlusion.
4.  Serial ECGs are critical but they must be scrutinized for changes, which may be very subtle (see below).

This was a 30-something woman with bilateral trapezius pain.
There are hyperacute T-waves
There is 1 mm of STE in V3, so the formula can be used to find if this is normal STE or not
QTc is 444 ms.
STE 60 V3 = 1.5 mm, R-wave amplitude V4 = 15 mm
Formula value is 23.1, which is close to being an LAD occlusion value of 23.4.  It is below the cutoff of 23.4, but above my safe value of 22.0

4-variable formula value = 19.38 (consistent with LAD occlusion)
The QRS amplitude in V2 is very small for early repolarization, and that is why the formula gets this right.

This was missed by the physicians, even with a bedside speckle tracking ultrasound: no wall motion abnormality was seen.

The first troponin I was below the level of detection.

She was admitted for rule out MI, but the 2nd troponin was elevated, so she had another ECG recorded: 
This demonstrates what normal T-waves look like for this patient, and that she has zero ST elevation at baseline.
Fortunately for her, the artery had spontaneous repefusion.  '


She went for angiogram immediately after this and had a thrombotic LAD that was open with TIMI-3 flow.  It was stented.

Case 10.

A 25 year old with Epigastric Discomfort, Worse Supine, Better Sitting Up.

STE60V3 = 2.5 mm
 computerized QTc = 437 
(notice how it lengthened from the earlier values of 372 and 402 ms!)
R-wave amplitude V4 = 9 mmQRS = 26
3 variable formula = 25.839 (greater than 23.4 is all but diagnostic of LAD occlusion)
4-variable formula = 19.0 (greater than 18.2 is all but diagnostic of LAD occlusion)

The cath lab was activated, and angiogram showed a 90% distal LAD lesion with thrombus.  The post PCI ECG showed reperfusion T-waves.  


  1. Hi
    Thank you for this Epic post, though i'm in ENT residency program i always troll around on this blog and learn so much, again thank you.

    1. I am impressed!! An ENT resident who is looking at this blog. Amazing. You must be a great doc to have such wide interests and be so curious.
      Steve Smith

  2. Great post! For case 1, would you need the formula given the inferior ST-T wave changes?

    1. I would diagnose it without the formula. To most interpreters, the inferior leads are pretty subtle, and they might not even see them. You don't "need" the formula if you recognize it as LAD occlusion. It is an aid. If I think the ECG represents LAD occlusion, I don't not let the formula dissuade me. I use it mostly to assess cases that I do NOT think are LAD occlusion, and I am occasionally surprised.

  3. Hello...
    In most of the cases, I have been looking at anterior leads and focusing on V4, comparing its R and T waves. And then i call it LAD occlusion if R wave height appears low comparitavely and it is.. Can we get such findings in normal variants as well?

    1. don't focus just on V4. The 4-variable formula is your best guide, becuase it incorporates the QRS in V2 also.

  4. How do you calculate the QTc ?
    Because the formula I use is the bazett formula and it is inconsistent with the calculated QTc in many cases for example in the first case the HR is 78 and the qt is 8 mm so by the bazzet formula the QTc is 365 but in the case you wrote 420
    Also in the second case
    The HR is 69 and qt is 9.5 mm so the QTc is 405 and you wrote 455 ??
    So how do you calculate it sir ???

    1. Here is one viewpoint for the answer to your question — — and Here is a CALCULATOR for the various formulas — — That said, I think ( = my opinion) that no formula is perfect — but what counts is: i) Consistency in calculation; and ii) The clinical context. For me ( = my opinion) — rather than specific numbers, it is more important to know WHY you are calculating the QTc for the particular patient at hand — and often in patients who you follow — you want to know TRENDS comparing the QTc with what was previously calculating on whatever circumstances apply For example, if using Sotalol at higher doses — generalities count, but seeing a progressively increasing QTc is what will clue you into perhaps needing to lower the dose rather than a specific number. LOTS depends on the clinical case at hand! Hope that helps — :)

    2. Hi ,
      a great post indeed , however , i'm not get any of the QTcs similar to what is being posted here , this leads to a different Formula values . Moreover , i have tried Fridericia method to calculate the Qtc not only Bazzet's , but it only got me a narrower Qtc . If you please would calrify how to get the actual Qtc in the above mentioned cases . Thanks for your efforts dear professors .

    3. The formula was developed with Bazett, but some of these ECGs use Hodges (which is why you are not getting the same nubmer). Hodges results in a shorter QTc. Bazett's would be more accurate. Hodges would be less sensitive and more specific.

  5. Hi,
    Given the hyperacute T waves in case 12, is the formula necessary?


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.

Recommended Resources