Thursday, December 27, 2018

Profound ST Elevation in V1-V3. What do you think?

This ECG comes from Pierre Taboulet ( an ECG whiz who codes a lot of ECGs for Cardiologs' Artificial Intelligence Deep Neural Network algorithm (

I do research on Cardiologs' algorithm:
Smith SW et al. A Deep Neural Network learning algorithm outperforms a conventional algorithm for emergency department electrocardiogram interpretation.

Cardiologs founder, Yann Flereau, was named by MIT as the European Innovator of the Year.  Wow.  What an honor.

This ECG is displayed on the Cardiologs platform.

What do you think?

Clinical History: This patient had no ischemic symptoms, but did have a systolic BP of 220 mmHg.

You can see that the platform provides several important numbers, but unlike any other platform that I know of, it also displays the results of the 3- and 4-variable formulas for differentiating early repol from LAD occlusion (upper right corner).

3-variable = 23 (less than cutoff of 23.4 for LAD occlusion)
4-variable = 17.1 (also below cutoff, this time less than 18.2 which is cutoff for 4-variable formula)

Using the following measurements:
STE60V3 = 3.0
QTc = 417
RAV4 = 16.5
QRSV2 = 18

I obtained: 
3-Variable = 22.8 (but if I measure STE60V3 = 3.5, then 23.41)
4-Variable = 17.7  (but if I measure STE60V3 at 3.5, then 18.26)

The algorithm did not diagnose LVH.  My immediate impression of the ECG was LVH.  But it does not meet any of the standard criteria.

However, it does meet new LVH criteria, which were recently derived and validated: sum of the amplitude of the deepest S wave in any lead plus the S wave in lead V4. If the deepest S wave is in V4, then double that value. If the total is greater than 2.3 mV (23 mm) in women and greater than 2.8 mV (28 mm) in men then left ventricular hypertrophy is diagnosed. 

S-wave is in V2 = 17 mm
S-wave V4 = 9 mm
Total = 26 (not greater than 28), so not LVH by the new rule!

Nevertheless, it has the look of LVH.

Peguero JG et al. Electrocardiographic criteria for the diagnosis of left ventricular Hypertrophy. JACC 69(23):1694-1703; April 4, 2017.  Still, sensivity was only 62%, with greater than 90% specificity.

The formulas were not derived or validated on ECGs showing LVH, and their accuracy in this circumstance is uncertain.

I do believe that the 4-variable formula will work very well in LVH.  But we need to test it.

My subjective interpretation, and that of Pierre's, is of LVH with secondary repolarization abnormalities, including pseudoSTEMI ST elevation in V1-V3.  But lead V2 has a worrisome amount of ST elevation, and in a chest pain patient, I would be worried about STEMI.

The Ratios of STE to S-wave:

V1: 2.5/16 = 16%
V2: 3/17 = 18% (this is a high ratio even for LVH)
V3: 2/15 = 13%

Here is a figure from K. Wang showing the normal ST/S ratio in LVH:

Bedside echo confirms marked LVH without anterior wall motion abnormality:

This was the patient's baseline LVH ECG.

My Comment by KEN GRAUER, MD (12/27/2018):
A warm welcome to our friend and colleague from France = Dr. Pierre Taboulet, who always offers the most challenging of tracings. As per Dr. Smith — I greatly enjoyed working with the Cardiologs’ ECG platform, and found it expeditiously helpful to have Dr. Smith’s formula values already calculated and readily displayed for me at the time I viewed the ECGs I was interpreting.
  • I would add the following comments to the ECG discussed in this case. For clarity — I’ve reproduced this ECG, to which I’ve made a few additions (Figure-1).
Figure-1: The ECG in this case, with R wave amplitude in lead aVL noted, and showing the mirror-image of the QRS complexes and ST-T waves in leads V1 and V2 (See text). 

Over the years, more than 50 criteria have been proposed in the literature for the ECG diagnosis of LVH. In general, sensitivity of these criteria by even the best of electrocardiographers is lacking — typically under 60%. The “good news” is: iDespite poor sensitivity for the ECG diagnosis of LVH — when certain criteria are satisfied, specificity for LVH may be excellent (ie, >90-95%); andiiEchocardiography is usually readily available, and provides an excellent accurate method for assessment of chamber enlargement. Additional points regarding the ECG diagnosis of LVH include the following:
  • The Clinical History helps in ECG assessment for LVH. It is important to appreciate that even before looking at the ECG itself — the chance of true chamber enlargement is greatly increased IF the patient is older, and has longstanding hypertension and/or underlying heart disease likely to predispose to chamber enlargement (ie, heart failure, cardiomyopathy, valvular disease, etc.).
  • Specificity of the ECG for LVH is greatly enhanced if, in addition to voltage criteria — the patient also manifests ST-T wave repolarization abnormalities consistent with LV “strain”.
  • The physiologic basis for the finding of LVH on ECG, is that with increases in left ventricular mass — electrical forces directed toward the LV (left ventricle) increase, and this produces larger R wave amplitude in one or more left-sided leads (ie, leads I, aVL, V4,5,6). As a corollary — right-sided leads (ie, leads V1,V2) often develop deeper S waves, reflecting increased forces moving away from the RV, and toward the hypertrophied LV. Most ECG criteria for LVH reflect interplay between relative R wave and S wave amplitudes in various leads.
  • Because of the overall poor sensitivity of the ECG for assessment of LVH — voltage criteria for LVH are not always satisfied in patients with clear Echo evidence of chamber enlargement. That said, one may still strongly suspect LVH from assessment of an ECG if: ithe patient is of a certain age (ie, patients under ~35 often manifest larger QRS amplitudes not indicative of true chamber enlargement)iithe patient has longstanding hypertension and/or other underlying heart disease predisposing to LVH; andiiiLV “strain” is seen despite the lack of voltage criteria. In such cases — my practice when interpreting ECGs for others has been to write, “Probable LVH” as my assessment.
  • CLICK HERE — for more on My Approach to ECG Assessment of LVH.

Regarding the ECG in this case:
  • Standard criteria for the ECG diagnosis of LVH are met in this tracing. Accepted voltage criteria for LVH include an R wave amplitude ≥12 mm in high lateral lead aVL — which is present in this case. NOTE: Although challenging to assess due to lead overlap — the clearly darker density of the overlap area in lead aVL to me suggests the peak of the R wave in lead aVL at the point I indicate by the horizontal RED line in Figure-1.
  • PEARL: I find this voltage criteria of ≥12mm in lead aVL especially helpful when the axis is leftward (I measure a frontal plane axis of -10 degrees for this ECG). As the highest “lateral” lead — aVL sees the vector of LV depolarization almost directly toward it, when the axis is leftward. Lower-lying chest leads may not show increased QRS amplitude in such cases.
  • Another (albeit less-used) standard LVH criterion that may be met in this case — is the Cornell criteria. These criteria state that LVH is present IF the sum of the R in aVL + S in V3 ≥20 mm (for a female patient) or ≥28 mm (for a male). We are not told if this ECG is from a male or a female — but if a female, I count 12 (R in aVL) + 15 (S in V3) = 27 mm.
  • Finally, we are unable to tell if the LVH voltage criterion using R wave amplitude in lead I + depth of the S wave in lead III (if this sum ≥25 mm) is fulfilled — because the R wave in lead I is cut off ...
  • The above said, even if voltage criteria were not met in this case — I strongly suspected LVH as the cause of the profound ST elevation in anterior leads because: iThe history is strongly in favor of LVH (No chest pain; marked hypertension); iiST-T wave changes consistent with LV “strain” and/or ischemia are seen in multiple leads; andiiiThe SHAPE of the ST-T wave in leads V1,V2 is much more suggestive of LVH than of acute STEMI. Explanation: Just as some patients with LVH preferentially manifest deep S waves in right-sided leads instead of tall R waves in left-sided leads — ST-T wave changes of LV “strain” may manifest a mirror-image” (reciprocal) pattern of LV “strain” in right-sided leads. Doesn’t the inversion of leads V1 and V2 shown in the INSERT of Figure-1 manifest an ST-T wave shape that would strongly suggest LVH if present in V5, V6?
  • PEARL: Clinical correlation is KEY in this case. All bets would be off if instead of no chest pain, this patient had worrisome new-onset symptoms. If this was the clinical scenario, then the ST segment coving in multiple leads, with symmetric T wave inversion + 2-4 mm of J-point ST elevation in leads V1-thru-V3 might well represent an acute cardiac event. But in a patient without any such symptoms — marked LVH in a patient with severe longstanding hypertension may sometimes produce very similar ST-T wave changes as are seen in Figure-1.


  1. Thank you, nice presentation! I think there's at least one "standard" voltage criterion positive for LVH: R aVL>11 (but there's overlap of the inscribed rows, so difficult to be certain).

    1. I think it is 10 mm, but as you say, difficult to know for certain.

    2. Another voltage criterion positive here is Cornell's: R aVL + S V3 > 20 (woman) or >24 (men).

      You mention about a new LVH criterion, very simple indeed. By the deepest S wave in any lead, do you mean any of the 12 leads or just out of 6 precordials?

      Thank you

    3. @ Unknown — I measure 12mm for R wave amplitude in lead aVL. I thought the "double darkening" supported that measurement ( = my opinion). I discuss this in more detail (and acknowledge the difficulty in measurement) in My Comment (above).

    4. To unknown: yes, just precordials

    5. @ Unknown — I already addressed Cornell Criteria, as well as another potential criterion above in my comment. As I mentioned, there are more than 50 criteria in the literature for ECG diagnosis of LVH — which to me means that we lack truly excellent ones (ie, a criterion with GOOD sensitivity AND specificity). The interesting point about this tracing, is that the most commonly used criterion (and in my opinion, the easiest and best to remember = deepest S in V1,V2 + tallest R in V5,V6 ≥35mm) is not satisfied in this case despite an echo showing such marked LVH — :)

  2. Dr Smith, you mention the STE to S-wave ratios, and that 18% (in V2, specifically) is high. What are normal STE/S ratios in LVH?

    1. No one has studies this. I tried to compare LVH with, and without, occlusion, but it was too hard to find cases with occlusion. I believe this is because when there is pre-existing LVH, the occlusion shortens the S-wave and eliminates the LVH criteria! Nevertheless, that should not keep me from measuring the normal ST/S ratio in patients with LVH. But in my experience, the ratio is virtually always less than 0.15 and the vast majority are less than 0.10. I will put in a figure showing the normals.

    2. @ Anonymous — As per Dr. Smith, R & S wave amplitude in certain leads may be affected in the setting of anterior infarction. Remember that the ECG just shows the "net effect" of all of the millions of vectors, oriented in virtually every direction ... For example, with LVH (and a resultant increase in lateral lead R wave amplitude) — this may oppose (reduce) early chest lead r wave height, and therefore almost mimic anterior infarction (producing QS complexes). With anterior infarction, there is loss of positive anterior forces — and by reducing “opposing forces”, this may alter anterior S waves and lateral chest lead R waves … It’s hard to predict how much of an effect this will have …

  3. May I ask you guys if can I use those new criteria of LVH by V4 lead for athlethes?

    As you know all ECG criteria for LVH so far for athlethes do not correlate with ECHO parameters.

    1. Piotr,
      As far as I know, it was not studied in athletes. Athletes, however, tend to have their voltage in R-waves, not S-waves.


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