Monday, November 27, 2017

A 54 yo male with sudden chest pain. Computer says normal. Paramedic disagrees.


There is now an Android app for the 3- and 4-variable formulas. It is of course free (#FOAMed).  It was written by Yannick Schäfer (a medical student in France):

Remember there is also an iPhone app called "SubtleSTEMI"


This was sent by a very astute paramedic.

A 54 year old male came to the door of the fire department because of sudden chest pain while working.  It was squeezing and substernal.

The medic recorded an immediate ECG:
What do you think?
There is ST elevation, but it looks exactly like normal ST elevation ("Early Repolarization"), right?
By the way, "Unconfirmed" means a human needs to overread it.

This medic wanted to be certain that this ST elevation with large T-waves was normal ST Elevation, and not a subtle LAD occlusion that only appears to be normal.

He recorded another 1 minute later:
Not much change.
The medic applied the LAD occlusion vs. early repol formula immediately.

1st ECG:
STE60V3 = 3.5
QTc = 390
QRSV2 = 19
RAV4 = 11

3-variable formula value = 23.61 (most accurate, but not most sensitive, cutpoint is 23.4)
4-variable formula value = 18.18 (most accurate, but not most sensitive, cutpoint is 18.2)

2nd ECG
STE60V3 = 3.5
QTc = 383
QRSV2 = 20.5
RAV4 = 10.5

3-variable formula value = 23.36 (lower)
4-variable formula value = 17.72 (lower)

The medic used the 3-variable formula and obtained values of 23.4 and 23.5 (positive)

He activated the cath lab from the field.

The cath team was ready when he arrived less than 5 minutes later.

Before going to cath, the patient had this ECG in the ED:
Not much change.
STE60V3 = 3.5
QTc = 450
QRSV2 = 19.5
RAV4 = 10
3-variable formula = 27.46 (very high)
4-variable formula = 21.49 (very high)
This last ECG obtains a much higher value because the computerized QTc measurement, at 450 ms, is much longer.  Even if we doubt the last QT measurement by the computer, and assume that it is much shorter, with a QTc a value of 400, both formula values remain very high.

The MDs in the department did not think it was an MI.

The patient went to cath within 5 minutes and had a 100% LAD thrombotic occlusion.

This was his ECG after stenting:
Now the EKG is normal (and the computer would agree!)
The ST elevation and tall T-waves are all resolved.
This would be how the patient's baseline ECG would have looked, if one had been available.
This reperfusion was so fast that the peak troponin was only 0.3 ng/mL.  There was no residual wall motion abnormality.  Symptom onset to balloon time was less than 30 minutes.

Learning Points
1. This shows how any individual patient's normal ST segments may have zero ST elevation.
2. Other individuals may have quite a bit of normal ST elevation.

Therefore, if there is any ST elevation, it is up to you (not the computer!) to determine if it is normal or ischemic.

The formulas are very helpful in this regard.

Again, the computer called the ECG "normal."  

I have argued that physicians should view these ECGs even if the computer interprets it as completely normal.  This is because the computer is so bad at finding subtle occlusions.  Physicians have argued that they don't have the time and that they will be no better at identifying these subtle cases than the computer will be.  

Well, a doctor might not see it, but a paramedic did.  Kudos!!

That is because the paramedic learned.  

I am sure that MDs can learn too!


  1. This is almost exactly how I presented when I called for EMS and I'm a 30 year Paramedic. I did not have "pain", just a slight pressure sensations across my chest with diaphoresis. EKG interpretation was just "abnormal", no suspected MI indicated. On arrival at the E.D. my CPKs were elevated slightly and troponin was barely elevated (E.R. Doc's words). Cardiologist came to evaluate me and suggested I go ahead and have the cath, could not definitively say I was having a cardiac issue. It turns out I had 95% occlusion of my LAD.

  2. Superb example of acute coronary occlusion without frank ST elevation. What caught my eye on the 1st tracing was the disproportionately tall T wave (compared to the QRS complex) in lead V3. And, in the context of this ST-T wave in lead V3 — the T wave in lead V4 looks to be a bit taller-than-it-should-be given corresponding R wave amplitude in this lead. These findings are subtle — but in a patient with new symptoms (and especially in conjunction with the formula scores of concern that are given here, despite fairly good R wave progression and a fairly unremarkable QTc interval) — these ECG findings should alert the provider to the need for careful scrutiny with frequent follow-up tracings. In the 1-minute-later tracing, the T wave in lead V4 now equals the height of the R wave in this lead, so these ECG findings in a patient with new chest pain is now clearly enough in this clinical setting to activate the cath lab. Dr. Smith’s formula provides further support. Dr. Smith notes in this blog post that “some physicians have argued that they don’t have the time” to review the ECG when the computer calls a tracing normal. The reality is that that it should take the “educated eye” no more than a few seconds to survey all 12 leads on an ECG — and that with regular practice, one’s skills can be easily honed (as they have been for the astute paramedic on this case). So, while true that the above ECG findings ARE subtle (and might pass for normal IF the clinical setting was “routine ECG done on a completely asymptomatic patient”) — these ECG findings ARE real, and in a patient with new chest pain need to be assumed as indicative of an acute event in progress until proven otherwise. THANKS to Dr. Smith for presenting this case.

  3. Hey there Dr. Smith!
    Would you support always activating the Cathlab as soon as we find a persistently altered 3 or 4 variable formula?
    I agree that the T waves look hyper-acutey, but I don't really see the "concerning" ST elevation in this one (is it V4? DII? Because we know that males >40yo can have "normal" 2mm ST elevation at the J point in V2 and V3. The formula shows 3.5mm at 60ms from the J point).

    And being honest, since the patient has good R waves in V2-V4 and a rather "short" QTc , I would've re-controlled this patient in the ED with further ECGs (BTW I think T-waves look taller on the second ECG), troponins, etc. I'm not sure I'd activated the cath lab right away.

    I'd appreciate your thoughts.

    Thanks in advance!

    1. That depends on other factors, but in general, I recommend intensive evaluation: serial ECGs, emergent high quality echo. That way you avoid false positives. On the other hand, if the pretest probability is high (sudden substernal CP in a 54 yo, then you can almost activate without an EKG!)

  4. As for me we need to evaluate clinical state (stenocardia, dyspnoe, diabetes, PAD, heavy smoker and any equivalents of cardiac dysfunction..) then we have STD with positive Smith's formula! We have young pt with nonsymptomatic with higher Smith's formula especially athlethes? Do stigmatic them for fullfilled ECG LAD occlusion criteria? I am not quite sure it is a good way, but symptomatic are under suspicion

    1. Yes, pretest probability is very important. See above.

  5. I think i would have missed this. I do pick up most coronary occlusions on this blog. Visually R wave progression/amplitude is good, but i guess i have to use the formula more liberally with hyperacute T-waves. Good work by the medic. Thanks for sharing.


  6. Fantastic case and hats off to my fellow paramedic. Can't express how happy I am to have a 4 variable formula available on my phone now!

  7. very interesting, and still a bit frightening for me, Steve. i could easily see myself convincing myself this is not a "STEMI" when this and twenty other ecg's are tossed in front of my face, done in triage for a variety of complaints, while i'm putting orders into the EMR for three septic patients.
    i guess it comes down to 1. is there a doubt? does this ecg look a bit funky? 2. am i concerned about early repol vs stemi equiv? if so, i just plug it into my steve smith iPhone app. and one ecg begets another. and sometimes an echo.
    thanks again, Steve.

  8. Except for the very helpful clinical context, I think I may have missed this one initially. On first glance the ST elevation and T-waves actually seemed to me to be at the bigger end of normal for such huge S waves and good R wave progression. Any comments on why the S waves are so much bigger during pain and then small after cath (and presumably at baseline)? Usually with LAD occlusion we would expect attenuation of the S wave approaching terminal QRS distortion. Fantastic job by this paramedic, and thanks for sending.

    1. I have the same question as you. Why does this patient have such big S waves during the pain?

    2. S-waves do not always disappear during LAD occlusion. If there are no S-waves in either V2 or V3, that is 100% specific for LAD occlusion vs. Early repol. But not even a majority of subtle LAD occlusion cases manifested this "terminal QRS distortion" in our study.

  9. This is one of the most subtle and hidden LAD occlusion I have ever seen and which I honestly would have not recognized (though I learnt to not undestimate every ECG and case). This is perhaps the most helpful use of your formula: it puts into our brain maximal attention for very subtle abnormalities. Many thanks for presenting this insightful case!

  10. Do you really need a formula to decide on cath lab activation here?? We activated a STEMI call recently in a chap with Chest pain and Hyper-acute T waves. Same outcome - cath lab activated and LAD occlusion found. I can see a little utility of formulas with prehospital staff however even they have access to their own cardiologist these days, atleast in Australia who can assess a 12 lead ecg faxed from the field. Do Cardiologists use these formulas?

    1. If you recognize this one as LAD occlusion and not early repol without using the formula, then good for you. In this case, no one but the medic thought the ECG showed anything but early repol. Cardiologists in particular do not recognize this ECG pattern as ACS. They need the formula more than anyone, but I doubt that many of them use it.


Recommended Resources