Tuesday, August 13, 2019

The ECG was correct. The angiogram was not.

This was sent by Cam Mosley and Michael Truax, LSU-Baton Rouge Residency.

74 yo male with previous MI years ago presented with chest pain and nausea.

An ECG was recorded:
Sinus rhythm with PVCs.
What else?

My interpretation (and theirs: they activated the cath lab): 

It is clearly a subtle proximal LAD occlusion (OMI).  

-- Subtle STE in I, aVL, V2
-- Hyperacute T-waves in aVL, V2, V3
-- Q-wave in V2, which is always abnormal
-- reciprocal ST depression in III, aVF.  Thus, the STE in V2 cannot be normal.

Here is the outcome he sent:

Cath negative other than 40% LAD stenosis.
They looked at aorta too and it was ok.
Initial troponin = 0.02 ng/mL (ref up to 0.030).

Here is my response:

"What did subsequent ECGs show?  Same?  Or resolution?  That is the key.  LAD can be briefly occluded with negative trops and no thrombus seen (the thrombus lyses).  But if this is the case, then the ECG will resolve. If the ECG is the same, then it is truly a false positive.  If the ECG resolves, then it was indeed what it looks like: LAD occlusion, though brief (transient)."

So he sent the post cath ECG:
Indeed, the ECG confirms reperfusion, with resolution of ST in I, aVL and V2, terminal T-wave inversion in aVL, full inversion in V2, and resolution of hyperacute T-waves.

As you can see, the angiogram is not the final arbiter of acute coronary syndrome.  Angiograms sometimes show no evidence of ACS.  The thrombus lyses and may not leave behind any visible culprit lesion.  Although unusual, it is not rare.

In this study, approximately 10% of Transient STEMI had no culprit found:

Early or late intervention in patients with transient ST‐segment elevation acute coronary syndrome: Subgroup analysis of the ELISA‐3 trial

One must use all available data, including the ECG, to determine what happened.

Final Diagnosis?

If the troponin remained under the 99% reference, then it would be unstable angina.  If it rose above that level before falling, it would be acute myocardial injury due to ischemia, which is, by definition, acute MI.  If that is a result of plaque rupture, then it is a type I MI.  The clinical presentation would be one of either unstable angina or type I MI, even if no culprit was found on angiography.

See this post for another case of OMI with a negative angiogram, and here are more references:

Inferior hyperacute T-waves and ST elevation. Angiogram is normal. What happened?

Bibliography, with edited abstracts

There were 821 cath lab activations and 86% were treated by mechanical revascularization. In 76 patients (8.5%), no coronary artery stenosis was documented. Observations documented angiographically included coronary spasms (6.6%) and muscle bridges (5.3%). During a mean follow-up of 11.2±6.4 months, one patient developed an acute myocardial infarction requiring coronary intervention. All other patients were free of any cardiac event.

Article 2, full text

Of 898 patients who had cath lab activations for primary PCI, normal coronary angiograms were obtained for 26 patients (2.6%). Among these, the diagnosis at discharge was a small myocardial infarction in seven patients (0.7%), acute (peri)myocarditis in five patients, dilated cardiomyopathy in four patients, hypertension with left ventricular hypertrophy in three patients, pulmonary embolism in two patients and misinterpretation of the electrocardiogram (ie, no cardiac disease) in five patients. Seven patients with small infarctions underwent angiography within 30 min to 90 min of complete relief of the signs of acute ischemia, and thus, angiograms during pain were not taken.   None of the 898 patients catheterized during ongoing symptoms of ischemia had a normal coronary angiogram. Spontaneous coronary spasm as the only cause (without underlying coronary atherosclerosis) for the evolving infarction was not seen. Thus, the causes of the seven small infarcts in patients with normal angiograms remain uncertain.

Characteristics of 690 consecutive patients with presumed STEMI referred for primary PCI.  87 (13%) had angiographically normal coronary arteries and were compared with patients with angiographically shown culprit lesions (control group; n = 594). Nine patients with significant coronary disease, but no identifiable culprit lesion, were excluded. Electrocardiograms (ECGs) from both groups were reviewed by 2 cardiologists blinded to angiographic findings.  On expert review of ECGs, 55% of patients in the normal coronaries group had ST-elevation criteria for STEMI (vs 93% in the control group, but the ECG was considered consistent with a diagnosis of STEMI by both observers in only 33% (vs 92% in the control group)   Left branch bundle block independently correlated with normal coronary arteries on multivariate analysis (odds ratio for STEMI 0.016).   The discharge diagnosis in the normal coronaries group was predominantly pericarditis (n = 72; 83%), but these were not adjudicated by the authors.  Other diagnoses were myocarditis in 3 patients (3%), Takotsubo cardiomyopathy in 2 patients (2%), presumed coronary spasm secondary to intravenous drug abuse in 2 patients (2%), cryptogenic AMI in 1 patient (1%), dilated cardiomyopathy in 1 patient (1%), massive pulmonary embolus in 1 patient (1%), cholelithiasis in 1 patient (1%), and pneumonia in 1 patient (1%).

The most likely alternative diagnosis suggested by both observers for the non-AMI ECGs in the normal coronaries group was normal variant ST changes (25% observer 1 and 26% observer 2) and early repolarization abnormality (25% observer 1 and 14% observer 2). 

The medical records of 941 patients undergoing coronary arteriography for presumed ACS within 48 h of onset were critically reviewed. In 70 patients (7.4%, 35 males) no CAD was documented. Alternative substrates of acute myocardial ischemia included coronary artery anomalies (7 patients, 10%), coronary spasm (10 patients, 14.3%), spontaneous coronary dissection (2 patients, 2.8%), paradoxical embolism through a patent foramen ovale (4 patients, 5.7%), embolism from left atrium or calcified aortic valve (4 patients, 5.7%), imbalance between oxygen demand and supply (20 patients, 28.5%), mitral valve prolapse (11 patients, 15.7%). No alternative substrates were found in 12 patients (17.1%). Absence of CAD is an uncommon finding in patients undergoing coronary artery angiography for ACS.

Comment by KEN GRAUER, MD (8/13/2019):
Superb case by Dr. Smith — for which the title tells all: “The ECG was correct; the angiogram was not”. As per Dr. Smith, there should be little doubt that in the setting of new-onset chest pain for this 74yo man with prior history of coronary disease — that the initial ECG ( = ECG #1 in Figure-1) represents acute OMI due to proximal LAD occlusion.
  • For clarity — I’ve taken out the long lead II rhythm strips, and have placed the initial ECG and post-cath tracing next to each other in Figure-1. I have done this because lead-to-lead comparison of these 2 tracings CONFIRMS each of my statements below.
  • I would add the following to the excellent discussion by Dr. Smith ...
Figure-1: The 2 ECGs in this case (See text).

MTHOUGHTS: Dr. Smith has noted abnormalities in 6 leads in the initial ECG ( = ECG #1). In my opinion — there are 11 (out of 12) leads in ECG #1 that are definitely abnormal the abnormal rhythm — which taken together, are diagnostic until proven otherwise of acute OMI.
  • There are 3 late-cycle (end-diastolicPVCs in ECG #1 (in simultaneously-recorded leads I,II,III; aVR,aVL,aVF; and V1,V2,V3).
  • Clinically, the finding of late-cycle PVCs (ie, PVCs that occur relatively later in the R-R interval— is of comparable diagnostic significance as the onset of AIVR, which is so commonly seen when there is reperfusion of the “culprit artery”. Retrospectively — we see that this is exactly what happened! That is, the acute abnormalities in ECG #1 have in large part resolved in ECG #2 due to spontaneous reperfusion of the “culprit artery” — which was heralded by development of these late-cycle PVCs!
  • As per Dr. Smith — Lead Vin ECG #1 is clearly abnormal because of: ithe Q wave that should not be present in lead V2, especially since there appears to be a tiny-but-present initial r wave in lead V1; iislight ST elevation of abnormal shape in lead V2; andiiithe hyperacute T wave in lead V2. That said — I suspect there is slight misplacement of the lead V2 electrode in ECG #1 — because it just doesn’t make physiologic sense to “sneak in” a qRs complex in V2, that falls in between the rS complex in V1 and the rS complex in lead V3. That said, regardless of slight lead misplacement of V2 in ECG #1 — the above-stated abnormalities are almost certainly still valid! NOTE: A much more logical sequence of R wave progression is seen in leads V1,V2,V3 in ECG #2 — which I believe supports my suspicion that there was some lead misplacement of lead V2 in ECG #1.
  • Considering the slight ST elevation and/or hyperacute T waves noted by Dr. Smith in leads I, aVL; V2, V3 — I thought that hyperacute waves were also present in leads V1 and V4-thru-V6. Normally, one should not see the 0.5 mm of ST elevation with frank upright T wave in lead V1. And — Aren’t the T wave peaks in leads V4, V5 and V6 broader (fatter) than they should be? NOTE: Confirmation that leads V1 and V4-thru-V6 clearly show acute changes in ECG #1 is forthcoming from lead-to-lead comparison with ST-T wave appearance of these same leads in ECG #2!
  • Finally — not only leads III and aVF — but also the 3rd inferior lead ( = lead II) show acute reciprocal changes in ECG #1. While the ST segment is not depressed in lead II of ECG #1 — the straightening of this ST segment in lead II, with abrupt angling to mark onset of T wave upslope in this lead is clearly abnormal, and consistent with acute change. NOTE:  Confirmation that ST-T wave appearance in lead II of ECG #1 clearly indicates acute change is evident from near normalization of the ST segment in this lead in ECG #2.
BOTTOM LINE: Dr. Smith makes the superb point in this case that, “The ECG was correct; the angiogram was not”.  As per Dr. Smith — skillful assessment of the post-cath ECG in context with the clinical history in this case tells us there almost certainly was transient proximal LAD occlusion despite the finding of no more than 40% narrowing of the LAD on cath. The main points embedded in my discussion above are:
  • The finding of late-cycle PVCs that then resolve as chest pain and hyperacute ECG changes resolve is further evidence supporting spontaneous reperfusion of a coronary artery that was transiently occluded.
  • The more leads you can identify that show abnormal acute changes — the more confirmation you have that an acute cardiac event is ongoing. Not all leads in ECG #1 show the same degree of acute abnormality. But when 11 out of 12 leads show what looks to be acute ECG changes in an older adult with new-onset chest pain — the inescapable conclusion is of an acutely evolving cardiac event.
  • The BEST way to hone your ECG interpretation skills is to BE SURE to look at follow-up tracings! While it might be easy to overlook the acute abnormalities in leads II; V1; and V4-thru-V6 — it is much easier to appreciate that these leads were abnormal in ECG #1, when you compare how different the ST-T wave appearance of these same leads is in ECG #2.
Our THANKS to Dr. Smith for this interesting case!


  1. One question for Dr Grauer. I can tell the PVCs' morphology is suggestive of ischaemia, but how do you tell they are "late cycle"? Are there any criteria we can refer to? Thanks!

  2. While I do not have a specific numeric definition for a “late-cycle” PVC — I would answer your question as follows: The rhythm most commonly associated with acute reperfusion is AIVR ( = Accelerated IdioVentricular Rhythm) — which is simply a ventricular rhythm that occurs at a heart rate faster than 20-40/minute (which is the typical idioventricular response rate) — yet slower than “VT” ( = Ventricular Tachycardia, which in its potentially lethal form is usually associated with a heart rate of ≥130/minute. I’ve considered ventricular rhythms at rates between ~110-130/minute as in a “gray zone” between “AIVR” (or “slow VT”) — and “fast VT”. The leaves us with a “usual rate” for AIVR as being beween ~70-110/minute. Not commonly appreciated — is that “late cycle” PVCs (sometimes called “end-diastolic” PVCs) are a comparable arrhythmia “marker” to AIVR, as suggesting the occurence of acute reperfusion when temporally associated with reduction of chest pain and ST elevation. The underlying sinus rhythm in ECG #1 in my Figure-1 occurs at a rate of ~80/minute (ie, corresponding to an R-R of just under 4 large boxes). NOTE that the COUPLING interval for each of the 3 PVCs seen in this tracing is ~3 large boxes. Given the relatively faster sinus rate — this places these PVCs as occuring in the LAST quarter of the R-R interval, which is later-than-expected — or to me, subjectively, a “late-cycle” PVC. Synthesizing the presumed sequence of events in this case (as I did in the 2nd bullet of “MY Thoughts” above) — I interpreted these relatively late-occurring PVCs as another marker of spontaneous acute reperfusion in this case.


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