Tuesday, November 17, 2020

A woman in her 60s with VFib arrest and no STEMI on her post-ROSC ECG.

Case submitted and written by Sonika Raj MD, with edits by Meyers and Smith 

This case shows how the COACT trial was fatally flawed and dangerous. (see references below)

A woman in her 60s with past medical history of type A aortic dissection status post endovascular repair and end-stage renal disease on peritoneal dialysis presented to the Emergency Department in cardiac arrest. She was found unconscious in her car, which had run off the road just outside the hospital doors. A bystander initiated chest compressions in the field and called EMS, who brought the patient inside. The first rhythm observed was ventricular fibrillation. She was defibrillated once and achieved ROSC. Approximately 20 minutes after ROSC, her ECG is shown below. No prior ECG was available at that time.

What do you think? 

There is poor quality with wandering baseline, and all attempts should be made to repeat it immediately with improved quality if possible. However, I would argue there are significant findings here that are highly concerning for acute OMI involving the inferior, posterior, and lateral walls. Despite the wandering baseline, there is the appearance of slightly large volume T waves in leads III and aVF, with reciprocal negative hyperacute T waves in aVL. Leads V2 and V3 have almost definite STD (if the baseline were slightly more steady I could say definite). Lead V6 appears to show slight STE. Together, each of these findings supports the others, and the likelihood that they are only due to baseline wander is low. 

Again, all attempts should be made to get a crystal clear baseline immediately, however we all know that, in real life, sometimes it is not possible to get a perfect quality ECG.

The ECG does not meet STEMI criteria, but especially in the context of a ventricular fibrillation arrest, is diagnostic of inferoposterolateral OMI. The cath lab was activated. The patient was given a fluid bolus for potential RV infarction with hypotension, and pressors to maintain goal MAP. 

Interventional cardiology evaluated the ECG but they were reluctant to take the patient to the cath lab since the ECG didn't meet STEMI criteria.

Approximately 60 minutes later another ECG was recorded: 

Leads II, III, and aVF have some of the most hyperacute T waves I've ever seen, as well as leads V4-V6. There is electrical artifact and I cannot exactly make out the J points, but it likely also meets STEMI criteria, but that is immaterial compared to how obviously diagnostic the hyperacute T waves are. This is inferoposterolateral OMI. Of course notice the STD maximal in V1-V2 that signifies the posterior involvement. 

Shortly afterwards the patient was found to have a 100% occlusion of her left circumflex, which was treated with thrombectomy and PCI. 

Here is her pre-intervention angiogram:

Here is her post-intervention angiogram:

She was admitted to the CCU and hypothermia was performed because she was not following commands after ROSC was achieved. 

Her initial high sensitivity troponin I (drawn during arrest) returned at 27 ng/L (limit of detection 5 ng/L, 99% URL 15 ng/L for women, 23 ng/L for men)

The second troponin I was drawn several hours later and returned greater than 50,000 ng/L. At that time we also have a CKMB of greater than 600 with a total CK of 3506. 

The patient survived but long term follow-up is unavailable.

COACT Discussion, by Smith:

The COACT trial was fatally flawed, and now has many cardiologists convinced that if there are no STEMI criteria, the patient does not need to go to the cath lab.

Lemkes JS, Janssens GN, van der Hoeven NW, et al. Coronary Angiography after Cardiac Arrest without ST-Segment Elevation. N Engl J Med [Internet] 2019;Available from: http://dx.doi.org/10.1056/NEJMoa1816897

Should all patients with shockable arrest be taken to angiography regardless of STEMI or No STEMI?

There has long been controversy about whether to take patients with a shockable rhythm without ST Elevation to the cath lab, and a recent randomized trial showed no benefitCoronary Angiography after Cardiac Arrest without ST-Segment Elevation (COACT).  This study had a fatal flaw: they did not keep track of all the "Non-STEMI patients" who were NOT enrolled, but instead were sent for immediate angiogram.  It was done in Europe, where the guidelines suggest taking all shockable arrests emergently to the cath lab.  So it is highly likely that physicians were very reluctant to enroll patients whom they suspected had Occlusion MI (OMI), even if they did not have STEMI. These physicians did not want a patient with an OMI that was not a STEMI to be randomized to no angiogram.  This strong suspicion is supported by their data: only 22 of 437 (5.0%) patients in this study had OMI.

What percent of shockable arrests without STE have an OMI?  

This large registry in Circulation 2010 reported that at least 1 significant coronary artery lesion was found in 128 (96%) of 134 patients with ST-segment elevation on the ECG performed after the return of spontaneous circulation, and in 176 (58%) of 301 patients without ST-segment elevation. 

5% vs. 58%!!  It is clear that there was signficant enrollment bias in COACT.

We at Hennepin recently published this study

Sharma et al. (with Smith and others) found that among patients with shockable cardiac arrest who had OMI, the initial and subsequent pre-angiogram ECG were only 75% sensitive for OMI, with similar specificity.

Sharma A, Miranda DF, Rodin H, Bart BA, Smith SW, Shroff GR. Do not disregard the initial 12 lead ECG after out-of-hospital cardiac arrest: It predicts angiographic culprit despite metabolic abnormalities. Resuscitation Plus [Internet] 2020;4:100032. Available from: http://www.sciencedirect.com/science/article/pii/S2666520420300321

Learning Points:

Serial ECGs are of paramount importance.

When the benefit of emergent reperfusion is maximal is when the troponin is minimal. Troponin has no role in the decision to perform emergent reperfusion when it is diagnosable by ECG as in this case.

Ventricular fibrillation cardiac arrest should prompt consideration of cardiac catheterization. This case is an example in which the post ROSC ECG does not show STEMI but is indicative of OMI nonetheless. These are the post arrest patients who certainly benefit from emergent cath.

Lead aVL is useful when evaluating for reciprocal changes in the setting of suspected inferior lead ischemia.


MY Comment by KEN GRAUER, MD (11/17/2020):


I learned about the concepts of inter- and intra-variability in ECG interpretation some 30+ years ago, during the time I was involved in studying the pros and cons of computerized ECG interpretation programs (Grauer KG et al: J Fam Pract 24:39-43, 1987; Grauer KGMarriott HJL, et al: J Am Bd Fam Pract 2:17-24, 1989 — among others). Appreciation of these concepts is critical for optimal ECG interpretation of acute coronary disease.

  • ECG Interpretation is far from a perfect art. There is often no “gold standard” to “prove” which interpretation is correct (and more than a single “correct” interpretation is often possible). I have seen estimates of up to 30-40% inter-observer variability among expert ECG interpreters. This means that IF you give the same tracing to a series of experts — there may be up to 30-40% variation on the interpretation of various ECG parameters (rhythm, conduction defects, ST-T wave changes, etc.).
  • There is also ~10-20% intra-observer variability among experts in ECG interpretation. This means that IF you give the same tracing to the same expert on 2 different occasions (say, separated by weeks to a month or two) without telling that expert how they interpreted the tracing the first time — that there will be up to 10-20% variation on the interpretation of findings the 2nd time, compared to what that expert said at the time of their initial assessment.
  • Of Note: Inter- and intra-observer variability has been found to persist — even when interpreters agree to use identical criteria. There may be less variability if clinicians shared common training — but even under the best of circumstances, ECG interpretation remains a science and an “art”, with less than 100% reproducibility.
  • The Good News — is that despite the above humbling reality of inter- and intra-observer variability — attaining expertise in clinical interpretation should allow experienced clinicians when given the same ECG with an adequate history to arrive at a similar overall assessment and management approach.

Applying the Above to Today’s Case:

Drs. Raj, Meyers, Smith and myself all independently assessed ECG #1 (which I have reproduced for clarity, together with the follow-up tracing in Figure-1).

  • We were all told that the patient was a 60-something woman with underlying medical problems who presented to the ED in cardiac arrest. Approximately 20 minutes following ROSC — ECG #1 was obtained.
  • All of us independently thought that ECG #1 was suboptimal technically — with significant baseline wander and artifact. Ideally — we all would have wanted a follow-up tracing as soon as could be possible — but this was not done.
  • That said — We all came to the same conclusion regarding our interpretation of ECG #1: In the setting of resuscitation from VFib cardiac arrest, ECG #1 strongly suggests acute OMI. That said, regardless of whether or not millimeter-STEMI-criteria are present or not in ECG #1 — the combination of abnormal ECG findings seen in this tracing is more-than-sufficient to indicate immediate cardiac catheterization.

Agreement on STEMI Criteria:

There is general agreement among experienced clinicians as to whether a given ECG satisfies millimeter-criteria for an acute STEMI or not. As a result — the phenomena of inter- and intra-observer variability in ECG interpretation rarely come into play.

  • Unfortunately — inter-observer variability for interpreting whether or not OMI is present (in the absence of satisfying millimeter-criteria for STEMI) is huge. This deficit in our clinical approach has been the basis for much of the research performed by Drs. Smith and Meyers.
  • Dr. Meyers has summarized conclusions of this research in his 17-minute talk that is accessible on our September 3, 2020 post — supported by literature in our July 31, 2020 post — ECG findings of which I summarize in the Table below in Figure-2, taken from My Comment at the bottom of our September 13, 2020 post.

Figure-1: The 2 ECGs shown above in today’s case (See text).

MY Thoughts on ECGs #1: As noted earlier — there is significant baseline wander + artifact in ECG #1. Nevertheless — it is possible to interpret this tracing, which was obtained ~20 minutes after ROSC following VFib cardiac arrest.

  • The QRS complex in ECG #1 is narrow. The rhythm is irregularly irregular at a rate in the 40s. This is inappropriately slow AFib.
  • There is low voltage. The QTc does not appear overly long. The frontal plane axis is about +50 degrees. There is no chamber enlargement.

Regarding Q-R-S-T Wave Changes:

  • There appear to be small Q waves in leads III and aVF.
  • R wave progression is normal, with transition (where the R wave becomes taller than the S wave is deep) occurring normally here between leads V3-to-V4.
  • Unfortunately, assessment of ST-T wave changes is difficult, because no more than 1 or 2 QRST complexes appear in any given lead due to the excessively slow rate. That said — the 1 complex we see in lead III, and the 2 complexes we see in lead aVF all suggest subtle-but-real ST elevation. We know this is likely to be real — because both of the ST-T waves that we see in lead aVL show reciprocal ST depression that looks like a mirror-image opposite picture of the changes we see in leads III and aVF.
  • Each of the ST-T waves that we see in leads V2 and V3 appear to show ST segment flattening with slight depression. Thus, ST depression is maximal in leads V2 and V3.
  • We only see one ST-T wave in simultaneously-recorded leads V4, V5 and V6 — but this ST-T wave segment in each of these 3 leads appears to show ST elevation, with a T wave in lead V4 that is probably hyperacute (ie, taller and fatter-than-it-should-be given height of the R wave in this lead).

ECG #1: Putting It All Together — Suspicious ECG findings are seen in no less than 8/12 leads in ECG #1. While possible that the suboptimal technical quality of this tracing might account for erroneous signals from a couple of leads — I find it hard to ignore suggestions from 8/12 leads in this patient who has just been resuscitated from VFib cardiac arrest. Taken together — the findings in ECG #1 are telling us why this patient had VFib.

  • Looking at the entries that I’ve listed in Figure-2 — our interpretation of ECG #1 should suggest OMI until you prove otherwise, because the 1st, 3rd, 4th and 5th bullets in this Figure-2 Table are satisfied.

MY Thoughts on ECGs #2: As always — it’s insightful to compare serial tracings, especially once clinical follow-up is available. We can learn a lot by going back to our interpretation of the initial tracing, to see what the early abnormal findings looked like in ECG #1, and what they became in ECG #2!

  • The rhythm in ECG #2 remains slow AFib.
  • As per Dr. Meyers, despite the electrical interference artifact in the limb leads — there are dramatically hyperacute ST-T waves seen in each of the inferior leads — with equally dramatic reciprocal ST depression in leads I and aVL.
  • In the chest leads — disproportionately tall and fat T waves (wide at their base) are seen in leads V3-thru-V6 ( = hyperacute T waves) — in association with ST elevation in these leads.
  • Frank ST-T wave depression is now seen in leads V1 and V2.
  • Putting It All Together: As per Dr. Meyers — the appearance of ECG #2 confirms acute infero-postero-lateral OMI.
  • Retrospectively — we can see the “beginnings” of the hyperacute ST-T waves in the inferior and lateral chest leads of ECG #1.

Figure-2: ECG findings to look for when your patient with new-onset cardiac symptoms does not manifest STEMI-criteria ST elevation on ECG (See text).

Final LEARNING Point: Our purpose in developing criteria for predicting OMI when millimeter-based STEMI criteria are not satisfied — is to hopefully minimize the potential for inter- and intra-observer variability when interpreting acute tracings.

  • Had the interventional cardiologist in today’s case considered the criteria I’ve listed in Figure-2 — the decision to activate the cath lab could have been carried out much sooner (which would have increased the chance for achieving an optimal longterm outcome).


  1. I agree that the initial ECG in this case should typically be treated with immediate reperfusion. Splitting hairs over whether there is 1 mm of ST elevation in the inferior leads seems short sighted, the clinical context and reciprocal changes inform the clinician as to the diagnosis.

    Several points of clarification as it relates to COACT. The French (PROCAT)registry data from the Circulation article reports a 58% rate of "significant CAD" which is a stenosis of > 50%. You suggest there is a 58% incidence of acute MI - but these are very different. In COACT 2/3 of patients had significant CAD which is numerically higher than the 58% in PROCAT. Additionally there were about 2/3 with ischemic ECG changes at admission in the COACT trial. The assertion of bias in enrollment is not supported by the data.

    In the immediate angiography arm one third had emergent PCI in COACT which is essentially the same as the PCI rates in the PROCAT circulation article (for those without ST elevation). I am cautious about reading a randomized controlled trial with great scrutiny and then not subjection registry data to the same scrutiny.

    The decision post arrest to take patient for cath is a team based approach and requires evaluation of more than just the ECG ( age, comorbidities, lactate, resuscitation time, bystander CPR, etc). The driver for mortality in post arrest patient shifts from the ACS itself to neurologic injury - This is most common cause of death and no amount of stenting will attenuate anoxic brain injury.

    1. All good points. Nevertheles, a 5% incidence of Occlusion remains a very low proportion among such patients. More importantly, we will never know who was excluded from the study because, contrary to standards of such studies, they did not characterize the group that was excluded. How many excluded patients had occlusion without STEMI? How many were taken for angiography even though they did not have STEMI (because the treating physician was not willing to enroll the patient)? We will never know.


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