Monday, February 5, 2018

Chest pain and a non-diagnostic ECG. No worries, right?

A middle-aged male complained of chest pain and called 911.  3 prehospital ECGs were recorded.

Here is the time zero ECG:
Just some non-diagnostic T-wave abnormalities  
T-wave inversion in V2 is a bit abnormal, and frequently seen in posterior MI, but not diagnostic. 
 There is tiny bit of ST depression in V3 and V4 which is nonspecific and could be normal.  
There is a large T-wave in lead II which is also nonspecific.  

Here is the repeat at t = 9 minutes:
Lots of artifact, but no apparent change

Here is the t = 25 minute ECG:
Again, no significant change except that the ST segments in V3 and V4, which previously showed some minimal ST depression, are no longer depressed.

The patient arrived in the ED and, before another ECG could be recorded, he had a V Fib arrest.

He was defibrillated, then taken to the cath lab and had a 100% circumflex occlusion.

He did well.

MI in the setting of a normal or nondiagnostic ECG:

This is very common.  Approximately 30% of MI have no diagnostic STE, STD, or T-wave inversion; in other words, they have no ECG findings specific for ischemia.  It is also true that  approximately 25-30% of complete occlusions do not have diagnostic ST elevation (but most have some evidence of ischemia!).

This is the reason we do not rely on the ECG to rule out MI!   Over the years, I have shown you many ECGs that show evidence of MI that might not be seen by everyone, or evidence of coronary occlusion that might not be seen by everyone.

This one has no clear evidence of occlusion, and no reliable evidence of MI at all (occlusive or non-occlusive).

Learning Point: patients can have MI with a normal or nondiagnostic ECG, and they can even have complete occlusion!

That said, I showed it to Vince DiGiulio and he wrote back:

Going into this I figured the ECG showed ischemia since you shared it, but I'm being honest that I didn't look at the outcome when I formulated my opinion (I sound like Ken). My thoughts were:
  • My pretest probability of ischemia is high since Steve Smith sent this. I consider it similar to the patient coming in with a good story and positive Levine sign.
  • There is slight ST-depression in V3 relative to the PR, but absolutely no ST-depression in V6. In my experience, when there is a decent pretest probability, that correlates highly with isolated posterior MI and is almost always due to LCx culprit. If the patient came in with belly pain I wouldn't pay it might mind, but typical CP and this ECG has me thinking LCx until proven otherwise.
I wasn't at all surprised to hear it was a 100% LCx.

The circumflex territory is the most "electrocardiographically silent" of the three epicardial arteries.  50% of circ occlusions do not show diagnostic ST elevation, but most do show some ST elevation less than 1 mm, or some ST depression.

Many circ occlusions result in only ST depression in leads V1-V4 [formerly known as isolated "posterior MI", but now the echo and MRI societies want to call them all lateral or inferior -- (a mistake if you ask me, but more on that later)].  If you do suspect occlusion but see no ST depression or elevation, try posterior leads V7-V9.  Remember that the voltage of the QRS and of the ST segment and T-wave are diminished by the impedance of the lungs between the posterior wall and the posterior leads.  Thus, data supports a cutoff of 0.5 mm of ST elevation (not 1 mm) when there is posterior ischemia.  Like all other distributions, millimeter cutoffs have very imperfect sensitivity and specificity, and do not capture the changes of ACO that can be seen by a skilled interpreter.

Also, posterior leads are not sensitive enough to rule out posterior MI; frequently patients with right precordial ST depression from acute posterior MI do NOT have any posterior ST elevation.  So if you make the diagnosis on the 12-lead, do not reverse your diagnosis based on absence of STE in posterior leads.

In any patient, if you strongly suspect ACS and the patient has continued, refractory pain, the European Society of Cardiology AND the ACC and AHA all recommend urgent (less than 2 hours) angiogram.  However, this makes it very easy to have false positives, as we know that many patients have ongoing noncardiac chest pain.  So this requires great clinical skills (or luck)!   A positive troponin, or stat echocardiogram, or continued serial ECGs (or a cardiac arrest!) may help to identify these patients.

It should be rare that you activate the lab in the absence of all of these!


  1. Wonderful case and wonderful discussion.

  2. Great teaching case and well-presented! A cardiology attending of mine during residency used to refer to aVL as the "eVL" ("evil") lead. A subtle finding that I noted was that the ST segments in aVL (and also eventually in Lead I) were flat and the T waves were very symmetrical. I think that at times not enough attention is given to those findings - especially in the scenario of a patient with credible chest pain.

    As for the renaming of "posterior" infarctions as "lateral," don't make the "old-timers" mistake of thinking "but that's not the way we always did it!" There is a difference between the anterior surface of the heart and the suface of the heart that is located anteriorly. Perloff made a mistake when he made the unwarranted assumption that the posterior surface of the heart aligned with the posterior surface of the chest (though that shouldn't detract from the very significant contribution he made to ECG diagnosis!). What we thought of as "posterior" infarctions really are located in the basal LATERAL wall of the left ventricle. The heart is normally rotated in the chest so that the basal lateral wall of the LV is facing - for the most part - posteriorly in the chest. The old "true posterior" area can still infarct, but it will be manifested around V3 or V4 (and not well seen, either). I have not seen any references that state the lateral infarctions should be called "inferior" MIs. Initially ALL inferior MIs were called "posterior" MIs. Then we began distinguishing between lower posterior ("diaphragmatic" or "inferior") MIs and upper posterior MIs (near the base of the heart). Those eventually became known as "true posterior MIs" and the rest were "inferior" MIs. The mistake was assuming that the heart was aligned with the same topography as the chest cavity. It isn't!

    I totally agree with doing V7 - V9 leads when the patient has chest pain and a non-diagnostic ECG. The LCx tends to irrigate areas of the heart that are not well-covered by the standard recording electrodes.

  3. Our THANKS to Dr. Stephen Smith for showing us this case — which is humbling and highly educational. Superb pearl by Vince DeGiulio. There is no denying the subtle J-point ST depression in lead V3 on the original tracing (plus as mentioned by Dr. Smith — suggestion of a taller-than-expected T wave in lead II, plus an unexpectedly small negative T wave in lead V2). That said — despite the “high pre-test probability” of an acute event (because this patient had chest pain, and because Dr. Smith posted this case on his ECG Blog) — I also thought the changes on these 3 tracings were non diagnostic … That said, it is good to appreciate that there IS slight serial change (ie, when one compares ECG #3 to ECG #1 — the T wave in V2 is no longer negative, and there is no longer any J-point ST depression in lead V3 DESPITE absolutely no change in lead placement in view of identical QRS morphology on both tracings). This brings home 4 points: i) As per Dr. Smith — Even complete acute coronary occlusion may not always produce diagnostic ECG signs; ii) As per Vince — subtle ST depression in V3, but no ST depression in V6 in a patient with new-onset worrisome chest pain suggests that there may be LCx occlusion until you prove otherwise; iii) I’ll add to Vince’s pearl, that in my experience — subtle change in lead II not explained by clear ST elevation in leads III and aVF in a patient with new-onset worrisome chest pain also suggests the LCx is more likely to be the “culprit artery” (and the T wave in lead II here IS taller-than-we-should-expect on all 3 tracings given very modest QRS amplitude in this lead); and iv) During my training I was always told if your accuracy in predicting acute appendicitis was greater than 80% — that meant that you were NOT referring enough acute RLQ pain patients to your friendly surgical colleagues. That’s because even the best diagnosticians on history and physical exam will not be able to predict with perfect accuracy which patients have acute appendicitis — so some patients who are appropriately explored may turn out to have a normal appendix. These percentages have of course changed with newer imaging techniques — but the corollary to acute chest pain patients remains — namely, that prompt cardiac catheterization IS appropriate on clinical grounds alone without diagnostic ECG signs in certain patients — even though some of those patients with turn out to have normal coronary arteries when the cath is done. This is one of the reasons caths are done — because sometimes you just need to PROVE whether this patient with chest pain who is in front of you, does or does not have not acute coronary disease. THANKS again to Dr. Smith for presenting this highly insightful case!

  4. thank you, gentlemen. interesting conversations on what appear on quick glance to be very benign. (and resulting in v-fib arrest)
    thank you


Recommended Resources