Sunday, May 26, 2019

A man in his sixties with chest pain at midnight with undetectable troponin

Written by Pendell Meyers

A male in his 60s with no known past medical history presented at midnight with chest pain over the past 3 hours. The pain started just after eating, and at first he thought it was "reflux," however he decided to call 911 after a few hours when it did not improve.

Here is his presenting ECG:
What do you think?

Here are the relevant findings:
Slight STE in V1
2.5 mm STE in V2
Slight STD in V4-V6
Definite STD in II, III, and aVF
Hyperacute T-waves in V2, and likely also in aVL

These findings are highly specific for LAD occlusion. We have many cases of this pattern on this blog, involving STE and hyperacute T-waves in V1-V2, with STD in the lateral leads (see below for links). In this case there is also reciprocal STD in inferior leads and hyperacute T-waves in aVL. 

Side note: Sometimes this pattern includes morphology which some have described as "new tall T-wave in V1," or sometimes "T-wave in V1 greater than in V6" which is essentially just a description of a hyperacute T-wave in V1, with the exception that it would be better described as both tall and fat/large/wide/bulky T-waves, not just "tall". Dr. Smith noted "new tall T-wave in V1" in 14% of cases of early repolarization vs. 34% of subtle LAD occlusions in the derivation study of the anterior OMI formula. This initial ECG does not have this pattern, however lookout for this pattern in the serial ECGs.

Our team recognized that this ECG represents LAD occlusion. But it does not technically meet STEMI criteria. We immediately called the interventionalist to explain our concerns. They evaluated the patient and were also concerned, however they felt that immediate cath was not warranted because the STEMI criteria were not met.

We began maximal medical therapy, including nitroglycerin infusion. 

To make matters more difficult, the first troponin of course returned undetectable (our contemporary assay rarely budges until 4-6 hours after onset of OMI).

During this time, we incessantly performed serial ECGs, trying to find one that convinced cardiology. 

Here is the clearest one we could get:

By my measurement, there is at least 1 mm in V1, and at least 2 mm in V2. This meets formal STEMI criteria for a male over age 40. The T-wave in V1 is now relatively hyperacute compared to the first ECG. Lead aVL is definitely hyperacute now that we have a clear tracing of it. ST depression in V4-V6 is more evident.

Despite the fact that we believed this ECG meets STEMI criteria, cardiology colleagues disagreed. This is not uncommon, and not unexpected, as prior evidence (see the end of the post for literature review) shows that we have very poor inter- and even intra-reader reliability for measuring the ST segment.

The cardiologist stayed in the room personally to see if the patients symptoms and ECG findings would be persistent despite medical therapy.

Sure enough, the ECG findings and ischemic pain indeed persisted over the course of 20-25 minutes, and the cardiologists correctly decided to proceed with emergent cath rather than further medical management, as is recommended by multi-national guidelines regardless of the ECG findings or interpretation.

We got one last ECG before the patient went to cath:
Now there is ST depression in V3, almost as if a de Winter's T-wave is developing.

 The time between presentation and cath was just under 2 hours.

100% mid LAD thrombotic occlusion was found and stented with excellent angiographic result.

Here is an ECG the next day confirming downstream reperfusion:
Of note, there is lack of pathologic Q-waves and expected reperfusion findings, suggesting relatively favorable infarct size compared to the at-risk territory.

Echo showed wall motion abnormalities of the anterior, septal, lateral, and apical areas. 

Peak troponin T was 1.87 ng/mL.

Learning Points:

We must be expert at subtle ECG findings of Occlusion MI, because our current paradigm is insufficient.

Serial ECGs can often help to make a difficult decision easier.

Persistent ischemia despite maximal medical management is highly likely to be due to Occlusion MI and is an indication for emergent catheterization regardless of ECG findings.

Contemporary troponin assays are frequently undetectable within the first 4-6 hours of Occlusion MI, when the benefit of emergent reperfusion is maximal.

See these other cases of LAD occlusion with similar subtle patterns of STE and hyperacute T-waves in V1-V2, with STD in V5-V6:

How long would you like to wait for your Occlusion MI to show a STEMI? Sometimes serial ECGs minimizes the delay.

McCabe et al, Journal of the American Heart Association. Physician accuracy in interpreting potential ST-segment elevation myocardial infarction electrocardiograms. Journal of the American Heart Association 2013;2:e000268.
A cross-sectional survey was performed by having emergency medicine physicians, cardiologists, and interventional cardiologists review 36 ECGs from the Activate-SF database of prospective STEMI activations. 12 (33%) of the 36 cases had no culprit lesion (defined as no STEMI), whereas the other 24 (66%) were true positives with total acute occlusion (defined as STEMI). This corresponded well with the actual overall rate of false positive STEMI activations in the entire registry of 36% which was recorded from prospective practice. For each ECG clinicians were asked, “based on the ECG above, is there a blocked coronary artery present causing a STEMI?” 124 physicians interpreted a total of 4392 ECGs. Overall kappa value of interreader agreement was only 0.33, reflecting poor agreement. Overall sensitivity and specificity for true positive STEMI (occlusion) was only 65% (95%CI 63-67%) and 79% (95%CI 77-81%). There was a 6% increase in the odds of successful interpretation with every 5 years of experience since medical school graduation. After adjusting for experience there was no difference in the odds of overall accurate interpretation between specialties. However, interventional cardiologists had the highest group specificity at 89%, while emergency medicine attendings had the highest group sensitivity at 74%. This excellent study is supported by many others showing poor inter-rater reliability (35-37).
Carley et al. What’s the point of ST elevation? Emergency Medicine Journal 2002;19:126-128.
Cross sectional study in which 63 clinicians who commonly prescribe thrombolytics for acute MI were asked to identify and quantify the degree of STE present in 3 sample ECG complexes. They were also asked to mark the ECG where they identified the J-point. Overall, STE was not identified in 23 (12%) cases. For figures 1-3 below, the percentage of doctors who correctly identified the J-point correctly was 29%, 61%, and 13%.

Tandberg et al. Observer variation in measured ST-segment elevation. Annals of Emergency Medicine 1999 Oct;34;448-52.
A blinded, paired-sample survey was administered to 52 subjects including emergency physicians, emergency medicine residents, and senior medical students. They were given a packet of 40 ECGs, blinded to the fact that it actually consisted of a random order of identical pairs of only 20 ECGs from patients with “enzymatically proven myocardial infarction.” They were asked to measure all ECGs, then the difference between the each reader’s two measurement of the same ECG was studied. The average difference in segment height among all groups was 0.28 mm. Overall statistical agreement between paired ST-segment measurements was very good (K=0.85). However, “one fifth of the time, intraobserver measurements of paired ST-segment elevations differed by more than half a millimeter.” When specifically asked whether the ST-segment elevation was greater than or equal to 2.0 mm, readers disagreed with themselves in 14% of cases.

Comment by KEN GRAUER, MD (5/27/2019):
Superbly (and very tactfully) written case by Dr. Pendell Meyers, on this previously healthy 60s male who presented with new-onset chest pain. My 2 questions on reviewing this case are:
  • Question #1: What will this Cardiologist do the next time he/she sees a similar case?
  • Question #2: Why will they do what they answered for Question #1 the next time?
For clarity — I’ve put together in Figure-1 the first 2 ECGs shown in this case — since the decision to cath was made before the 3rd tracing was obtained. 
  • NOTE: ECG #2 was not the 2nd tracing done in this case — because multiple ECGs were being frequently done after ECG #1 in an attempt to convince cardiology to take this patient to emergent cath ...
Figure-1: The first 2 tracings obtained from the patient in this case (See text).
Dr. Meyers convincingly restates the case that has been made many times in this blog:
  • The current paradigm for recognizing is insufficient — because too many cardiologists remain “stuck” on requiring millimeter definition of acute STEMI.
  • Reliance on this “millimeter definition” is potentially harmful to the patient — because emergent cath is indicated for persistent symptoms and clear evidence of acute ischemia. Time is muscle.
  • Despite literature documentation of poor inter- and intra-reader reliability among even expert clinicians — the ECG findings in ECG #1 should not be negated.
  • Reliance on contemporary troponin assays is a faulty strategy — because these won’t be elevated until valuable time has passed.
  • On the other hand — Stat Echo during symptoms can provide immediate confirmation that the findings in ECG #1 are real and merit immediate cath!
To this cardiologist’s credit — he/she did stay in the room and personally monitored the patient while medical therapy was being given — and total time from ED presentation until cath was under 2 hours.
  • That said — it should have taken less time to convince cardiology that new-onset chest pain in a previously healthy man in his 60s who presents with ECG #1 should immediately go to cath.
  • ECG #2 — clearly shows progression of ischemic findings, and should definitely have been enough if there was reluctance based on ECG #1.
  • SOUL-Searching — Acute medicine is an incredibly challenging profession. The reality is, that we have to learn from mistakes. None of us are flawless. I learned these sometimes very difficult lessons as an Attending charged with training and supervising residents over my 30-year career in academics. My only hope is that the cardiologist(s) involved in this case spent some time after cath seriously soul-searching what he/she might have done differently in this case — and what he/she will do the next time they encounter a similar situation.
MThoughts on ECG #1:  Dr. Meyers superbly highlighted the KEY findings (above). I’ll add the following thoughts:
  • The ST elevation + hyperacute appearance of the ST-T wave in lead V2 of ECG #1 should be recognized from 5 feet away. Calling the T wave in this lead “disproportionately tall” does not do it justice.
  • We know lead V2 is not an “aberration” — because of the distinctly abnormal shape of the ledge-like (straight) ST segment shape in neighboring lead V3. (Although more subtle — the T wave in V3 is also disproportionately tall-with-wider-than-it-should-be-base given the small r wave in this lead.).
  • Many other leads show ST segment flattening/straightening (V4, V5, V6) — with frank ST depression in leads II, III and aVF.
  • As per Dr. Meyers — there is a subtle-but-real hyperacute appearance to the ST-T wave in lead aVL.
  • Beyond-the-Core: If you look at the PVC in lead aVR — I believe the coved initial part of the ST segment in this lead is distinctly atypical, and reflects acute ST elevation.
Bottom Line re ECG #1: In a previously healthy patient with new-onset chest pain — One can’t discount the acute appearance of the ST-T wave in lead V2. That this is real is amply evidenced by ST flattening and/or depression in multiple other leads.
  • Regarding ECG #2 — The main difference I see between ECG #1 and ECG #2 is that there is now clear ST elevation in lead V1. I understand the tendency to try to convince oneself that “there is only 1 lead that has changed” — but this is a serial change in a patient with ongoing chest pain. Careful lead-to-lead comparison of QRS morphology in the 6 chest leads of both tracings confirms that chest lead electrode positioning has not appreciably changed. Therefore, the clear new ST segment elevation with hyperacute appearance now seen in lead V1 of ECG #2 is real.
  • My only hope is that the cardiologist(s) involved in this case takes another look at these 2 ECGs — and then compares this to the 3rd ECG in this case as a reminder of what happens when you wait just a little bit longer ...
Our THANKS to Dr. Meyers for highlighting events in this case.


  1. Hi there, is there any reason why not to use the 4-variables formula in this case? According to my calculation, the first ECG gets 24.3 points - highly suggestive for LAD occlusion. Thanks for the great case!

    1. In our study, we excluded obvious MI. One of the criteria for obvious was inferior ST depression. Here we have inferior ST depression, which makes normal variant STE nearly impossible and MI almost certain. If the formula were negative, it could not be trusted.

  2. Hi Dr smith, as you mentioned there are hyperacute t waves in aVL, so shouldnt the occlusion be at proximal LAD rather than mid LAD?

    1. No rule is 100%, not even close!

    2. The same thought occurred to me as well. However, though learning the anatomy of the different coronary arteries and the expected findings on the 12-lead ECG after an occlusion at some level is something that anyone striving to interpret ECGs at an advanced level should do - you must be prepared to accept the fact that anomalous anatomies are not rare. In this case it is possible that D1 was not the dominant branch off the LAD but perhaps a more distant D2. And perhaps S1 was not the dominant septal perforator. And even though, as Steve said - you can never be 100% certain - you should continue to make your decisions about which is the culprit artery based on the classical anatomy. After all, it IS correct most of the time - just not ALL the time.

      Jerry W. Jones, MD FACEP FAAEM


Recommended Resources