Wednesday, April 29, 2020

Cardiac Arrest. What does the ECG show? Also see the bizarre Bigeminy.

A 60-something woman presented after a witnessed cardiac arrest. CPR was started immediately.  EMS arrived and found her in a wide complex PEA rhythm. She was given 3 mg IV epinephrine and multiple rounds of ACLS over approximately 20 minutes.

Her husband stated that she had not been feeling well in the past 2 weeks and c/o dizziness as well as diarrhea.

She was never defibrillated.

I was texted this ECG in real time, but it turns out to actually be the 2nd one recorded in the ED.

What do you think?

This is what I wrote:

This looks like pseudoSTEMI to me.  What appears to be ST elevation seems to be a wide QRS.  I am only looking on my phone though.

Indeed, if you find the end of the QRS, which is easy to see in V1, then draw a line down to lead II across the bottom, then you can find the end of the QRS in lead II.  The go to the left and find that same point under aVR/aVL/aVF and under I, II, III, and you can draw this line back up to find the end of the QRS in these leads.

Sure enough, all of what you might think is ST Elevation is really QRS:

The K returned at 2.7 mEq/L.  This is commonly found after epinephrine for cardiac arrest, but could have been pre-existing and a possible contributing factor to cardiac arrest.

A recent similar case:

A 40-something with chest pain. Is this inferior MI?

Just as interesting is EKG 1, 24 minutes before the first:

What do you think here?

There is a bigeminy with very Bizarre looking PVCs.  These are frequently found in cases of Torsades.

Below, I post 4 more examples of ECGs that were recorded from patients who had Torsades.

1 hour later, this was recorded:
Now there is a very long QT.
See especially lead V2, where the QT is 640 ms, and QTc would be about 680 ms.
Interestingly, the computer measured the QT as 245 ms!!

Of note, she's on some QT prolonging medications as well as Hydroxychloroquine for Rheumatoid Arthritis, but no other known QT prolonging medications.

This was recorded at 6 hours:
Again, bizarre bigeminy with very long QT of both native beat and of the PVC

Computerized QT = 430
QTc = 449
Verified by overreading cardiologist

Obviously much longer, or they are U-waves

Potassium level

Troponin I peaked at 0.875 ng/mL, so an angiogram was done.

--Minimal coronary atherosclerosis
--No obstructive epicardial coronary artery disease or evidence of plaque rupture noted to explain prolonged QT or ventricular fibrillation cardiacarrest, suspect nonischemic mechanism

The estimated left ventricular ejection fraction is 45 %.
No wall motion abnormality.

Clinical Course

Initially, her potassium level was low which was replaced.  Despite of replacing potassium, she continued to have ventricular ectopy with prolonged QT interval and had 2 episodes of R on T phenomena followed by torsades de pointes, which was cardioverted.

Therefore, she underwent temporary pacemaker placement and overdrive pacing at a rate of 90 bpm to keep the heart rate up in order to prevent these PVCs triggering ventricular arrhythmia.

The ultimate reason for the long QT was never definitively determined.  Hypokalemia was unlikely because she continued to have ventricular arrhythmia despite of correcting electrolytes.  Drug-induced QT interval cannot be completely ruled out, but the tox consult found the she had definitely not overdosed and did not believe that therapeutic doses would do this.

Final Diagnosis: Cardiac Arrest due to Torsades from long QT of unknown etiology.


Below, I post 4 more examples of ECGs that were recorded from patients who had Torsades, either shortly before, during, or after.

3 of the 4 have similarly bizarre PVCs.

MY Comment by KEN GRAUER, MD (4/29/2020):
Cardiac Arrest with Bizarre PVCs/Torsades de Pointes: Intriguing case with many interesting features. I’ll focus my comments on selected points that enhance those superbly presented by Dr. Smith.
  • Relevant REVIEW of the CASE: The patient was a 60-something woman who presented in PEA cardiac arrest. Initial serum K+ = 2.7 mEq/L, which later decreased to 2.0 mEq/L — before treatment that returned it to normal. The patient was on some QT-prolonging medications. Cardiac cath showed minimal disease. The patient developed Torsades de Pointes, which was converted electrically, with recurrence prevented by overdrive pacing. She recovered. Final diagnosis was cardiac arrest due to Torsades from a prolonged QTc of uncertain etiology.
  • Dr. Smith showed a total of 10 ECGs in his comments above. For clarity — I’ve numbered the 3 ECGs I discuss according to the sequence in which they appear above in Dr. Smith’s comments (Figure-1):

Figure-1: The 3 ECGs I discuss, numbered according to their sequential appearance in comments by Dr. Smith (See text).

ECG #4 — This is the 3rd ECG that was done in this case (obtained ~1 hour after arrival in the ED).

QUESTIONS Regarding ECG #4:
  • In general — WHICH of the 12 leads on an ECG should be used to measure the QT interval?
  • WHY does this matter in a case like the one discussed today?
  • For ECG #4  WHICH of the 12 leads should be used to measure the QT interval?
  • The computer measured the QTc to be 245 msec. Since computerized ECG interpretations are supposed to be accurate for measurement of intervals — HOW could the computer be so off?
  • Extra Credit: WHAT is the rhythm in ECG #4? (Feel free to look ahead to ECG #5 when checking your answer.)

ANSWERS Regarding ECG #4:
  • The BEST lead on a 12-lead ECG for measuring the QT interval — is THAT lead in which you can clearly see the beginning and end of the QT — and, in which the QT interval is longest.
  • The reason this matters in this case, is that this patient was predisposed to, and later developed Torsades.
  • Assessment of the QT interval in ECG #4 is extremely challenging! This is because none of the 12 leads really show a distinct end point for the QT interval. One might think BLUE arrows in leads II and aVF indicate the end of a normal QTc. Other leads (ie, I, III, V3-thru-V6) manifest an even less distinct end point of the QT interval. But as per Dr. Smith — the KEY lies with focusing on lead V2, with the PURPLE arrow highlighting a shallow-but-distinct negative T wave. Thus, the QTc is markedly prolonged (ie, over 600 msec). BOTTOM Line: Measure the QTc in THAT lead in which you most clearly see the end of the interval, and in which the QT is longest ( = in lead V2 in for ECG #4). Be SURE to look at all 12 leads when measuring the QTc!
  • In general — computer measurement of intervals is accurate. BUT — computer measurements depend on where the computer determines the beginning and end of the interval being measured. As emphasized in the previous bullet — the end of the QT interval in ECG #4 is indistinct, and extremely difficult to determine.
  • PEARL #1 — IF it is difficult for YOU to determine the end of the QT interval, then it will also be difficult for the computer! In such cases — Be aware that you can not trust the computer estimation, and YOU need to assess the QTc yourself!
  • Extra Credit Answer: Did YOU notice that the P wave in lead II of ECG #4 is not upright? (P wave within the PURPLE circle). This means that the rhythm in ECG #4 was not sinus! However, the P wave in this tracing was positive in leads I and aVL (P wave within the BLUE circle). This means there was a low atrial rhythm. Confirmation that the rhythm in ECG #4 was not sinus — is forthcoming from inspection of lead II in ECG #5, in which the P wave now is upright in lead II (P wave within the RED circle).

What is the Difference between PMVT and Torsades?
A useful classification of WCT (Wide-Complex Tachycardia) rhythms — separates them into those that are monomorphic (with similar QRS morphology during the tachycardia) vs those that are polymorphic (in which QRS morphology varies).
  • When QRS morphology of an obviously irregular ventricular tachycardia changes from one beat to the next — the rhythm is called PMVT (PolyMorphic Ventricular Tachycardia).
  • Torsades de Pointes — is defined as PMVT that occurs in association with a long QTc on baseline ECG. This is especially true when the rhythm in question manifests the shifting QRS polarity around the baseline (ie, “twisting of the points”) that is characteristic of Torsades.

PEARL #2 — Distinction between PMVT vs Torsades is more than academic. Both treatment and the response to therapy tends to be different with these 2 entities.
  • Torsades — often has a multifactorial etiology (ie, drug-induced, electrolyte depletion, CNS disturbance and/or other underlying disorder that may predispose to QT lengthening). KEY aspects of treatment include IV Mg++, often given at high and sometimes repeated doses (even if initial serum Mg++ levels are not low). Optimal treatment of Torsades entails finding and “fixing” the cause of the long QTc. As was seen in this case — defibrillation and/or overdrive pacing may be needed.
  • In contrast — PMVT without QT lengthening most often has an ischemic etiology. Although IV Mg++ is also indicated as initial treatment of PMVT with a normal QT — it is clearly less likely to respond, than when the QT interval is prolonged. Instead, antiarrhythmic drugs such as amiodarone or ß-blockers may be needed — and/or treatment targeted to correcting ischemia.

ECG #5 — This is the 4th ECG done on the patient in this case (obtained ~6 hours after arrival in the ED).

QUESTION Regarding ECG #5:
  • WHAT ECG findings in ECG #5 predispose to development of Torsades?

ANSWER Regarding ECG #5:
  • Among the findings in ECG #5 that predispose to development of Torsades are: i) underlying bradycardiaii) a very long QTc interval (the GRAY arrow in the long lead II rhythm strip shows the end of the QT interval, which appears to be at least 600 msec)iii) Frequent and bizarre-looking PVCs, that also manifest an extremely long QT interval (the PINK arrow showing the end of the QT for the PVC, which is beat #8).
  • PEARL #3: As per Dr. Smith — bizarre-looking PVCs are often seen in patients who go on to develop Torsades.

ECG #8 — This is an ECG obtained from another patient with known Torsades, used by Dr. Smith for illustrative purposes:
  • There is atrial pacing of narrow-complex beats #1, 3 and 8.
  • Note again the presence of bizarre-looking PVCs ( = beats #2 and 4).
  • Note how difficult it is to assess the QTc of the narrow beats — because we really do not see a distinct end point to the T wave. More than any other lead — lead I suggests the QTc to be prolonged — but, the marked dip in the baseline seen in simultaneously-recorded leads II and III suggest this may be artifact.
  • Beginning with beat #2 — a run of Torsades is triggered. Note the characteristic shifting polarity of QRS complexes from positive-to-negative-to-positive-then-negative again (ie, “twisting of the points”).
Final Learning Points — It is common in practice when PMVT is encountered not to know (or be able to identify) IF the baseline QTc is prolonged or not. This is the case in ECG #8 — as I would not be certain that the baseline QTc is prolonged from this single tracing alone. As explained above — this distinction between Torsades vs PMVT with a normal QTc is important — because IF the QTc is indeed prolonged, then the PMVT is Torsades — in which case treatment and the response to therapy may differ than if the rhythm is PMVT without QT prolongation.
  • Hopefully either a baseline ECG and/or additional monitoring of the patient will reveal more clearly if the baseline QTc is prolonged.
  • Careful assessment of clinical factors potentially predisposing to one or the other etiology (Torsades vs PMVT) may help if the only ECG available does not clearly reveal the end of the QT interval in any of the leads (ie, Was serum K+ and Mg++ low? Use of QT-prolonging drugs? Acute ischemia? etc.).


Saturday, April 25, 2020

Another Shark Fin. With a twist.

I was reading stacks of ECGs for a study, without any clinical information.

I came across this one and immediately recognized it and knew the diagnosis (Pendell did too when I sent it to him):
There is a Shark Fin!

What is the diagnosis?

But this is not the kind Shark Fin we usually see, which is due to STEMI! 

This ECG is pathognomonic for severe hyperkalemia.  Wide QRS, large R-wave in aVR, Brugada-like ST Elevation in V1-V3 with inverted T-waves, extremely peaked T-waves (in many leads).

I went to the chart to find the case:

56 y.o. type 1 diabetic presented for evaluation of hyperglycemia. He had stopped taking insulin 5 days prior.  He reported SOB, cough, nausea, and vomiting.

K was 8.4 mEq/L.

pH of 6.96
bicarb 6
pC02 27. 
AG of 32. 

He was immediately given 10 u regular insulin and albuterol for shifting his hyper K. He was started on IVF with pressure bags. 50 mEq of bicarb was given for his severe acidosis. 7u/hr regular insulin drip was started.  

After 30 minutes, repeat labs showed no change in pH or bicarb level. K had improved to 6.9,

Here is the repeat ECG at K 6.9 mEq/L:

And this after K is down to 4.8 mEq/L:

Here are a few other cases of hyperkalemia with PseudoSTEMI:

Here are several cases of Shark Fin due to STEMI

MY Comment by KEN GRAUER, MD (4/25/2020):
GREAT case of “Shark Fin with a Twist”. I’ll suggest a few additional points to those made by Dr. Smith.
  • For clarity — I’ve put the 3 tracings sequentially together in Figure-1:

Figure-1: Sequential compilation of the 3 tracings in this case (See text).

As per Dr. Smith — the initial ECG in the ED ( ECG #1) shows:
  • QRS widening.
  • A large R wave in lead aVR.
  • Brugada phenocopy ST elevation in the anterior leads.
  • T waves that are pathognomonic for hyperkalemia in multiple leads (especially in the inferior and lateral chest leads).
  • PEARL — What I have found most helpful to facilitate instant recognition of hyperkalemia when there is significant QRS widening — is that not only are T waves in many leads tall and peaked — but that these T waves are symmetric (comparable T wave ascent and descent), and these T waves display an extremely narrow base. This will usually be more obvious in some leads than others — but it is unmistakable in leads V4, V5 and V6 of ECG #1.

Additional Features to Note in Figure-1:
  • Marked right axis deviation (predominantly negative QRS in lead I).
  • A relatively long PR interval (if anything, the PR interval is usually shorter with tachycardia).
  • A relatively longer QTc — with a longer-than-expected isoelectric ST segment best appreciated in leads V4 and V5 (which in the setting of hyperkalemia — suggests concomitant hypocalcemia).
  • A pseudo-infarct pattern is also seen in the limb leads (coved ST elevation in lead aVL with reciprocal ST depression in each of the inferior leads).

Sequential Changes — I found it of interest to follow over the course of these 3 sequential ECGs how each of the above additional features resolves as serum K+ normalizes:
  • There is no longer right axis deviation in ECG #3 (the axis is vertical at about +80 degrees, as the QRS is more positive than negative now in lead I).
  • Despite no reduction in heart rate — the PR interval has normalized in ECG #3.
  • The QTc is shorter, without any isoelectric ST segment.
  • The pseudo-infarct pattern in the limb leads has essentially resolved.

Additional Learning Points — The ECG presentation of hyperkalemia is multifaceted. In addition to tall, peaked and pointed T waves + QRS widening + a Brugada-like pattern in anterior leads — there may be:
  • Marked right axis.
  • Associated electrolyte abnormalities.
  • AV conduction defects.
  • Additional pseudo-infarct patterns.
  • Final Point You never know what the underlying ECG will look like until you correct the hyperkalemia and repeat the ECG. This is because whatever ECG findings may have been present before serum K+ increased (ie, ST depression or elevation) — may be masked by the QRS widening and tall, peaked T waves of hyperkalemia. In this case — ECG #3 showed little more than sinus tachycardia and some nonspecific ST-T wave changes.

Friday, April 24, 2020

Guess the culprit with ST Elevation in posterior leads

A middle aged man had off and on chest pain for 2 weeks, then 2 hours of more severe and constant pain.

Here was his ED ECG, which was identical to the prehospital ECG.  He did not get prehospital activation.

What do you think?

There is sinus rhythm with ST depression in I, II, aVF and V2-V6.  It is maximal in V3 and V4.  This usually means posterior MI, whether the T-wave is upright or not.

There is also some ST elevation in aVR, which must be present whenever there is ST depression in I and II (a lead between I and II is (-) aVR, opposite aVR; if ST segments in I and II are negative, then they must also be negative in the lead that is between them [(-) aVR].  If (-) aVR has ST depression, then aVR must have ST elevation.

ST Elevation in aVR is not the same as STE in other leads because there is no underlying myocardial wall; there are only atria!  So STE in aVR is reciprocal to ST depression elsewhere.

ST vector

In this ECG, the ST vector is both posterior (away from V3 and V4), upward, away from aVF, and rightward (away from I, II, V5, V6 and towards aVR.  It is mostly posterior, as the greatest ST depression is in V3 and V4.

When the ST vector is primarily posterior, the diagnosis is usually posterior STEMI.

I just read Ken's comments before publishing.

He wrote below that there is 2 mm of STE in aVR.  I did not and do not measure it that way.

So I went back to the original ECG and magnified it:

I put the top of the line just under the QRS onset, and the J-point, for both complexes

As you can see, the STE in aVR is indeed 1 mm, not 2 mm
The ST depression in V3 and V4 is much more pronounced than the STE in aVR.
Contrary to what Ken stated, the ST vector remains mostly posterior

What about subendocardial ischemia?

Subendocardial ischemia results in ST depression, but unfortunately, and rather mysteriously, it does not localize to the ischemic wall.  This has been shown in multiple studies of stress testing, in which the location of the ST depression does not correlate with the subsequent angiogram.

So if there is subendocardial ischemia due to an LAD lesion, the ST depression is not usually in the "anterior" precordial leads.  It is more often in II, V5, and V6.  ST depression due to a circumflex (lateral wall ischemia) or RCA (inferior subendocardial ischemia) similarly will not reliably localize.  It is possible that there will be ST depression in those locations, but not necessarily.

Conversely, if there is ST depression in the inferior leads, it does not mean there is "inferior ischemia;" it is far more likely to be reciprocal to ST elevation due to subepicardial ischemia (transmural, often extremely subtle) in aVL.

Similarly, STD in aVL is usually reciprocal to inferior ST elevation, not "lateral ischemia."

Likewise, ST depression in V2-V4 is most commonly reciprocal to ST elevation due to subepicardial (transmural) ischemia of the posterior wall.

ST depression maximal in V5 and V6 cannot be reciprocal to subepicardial, transmural ischemia under aVR because, as stated above, there is no ventricular myocardium beneath that lead, no STEMI under aVR.

Thus, ST depression maximal in V5, V6 is subendocardial ischemia.

Clinical Course 

I activated the cath lab.

Out of curiosity, I recorded posterior leads (Here, V4-V6 are really V7-V9 on the back):
V3 = V7, V4 = V8, V6 = V9
Notice the voltage is very small.
This is because the electrical impulse is impeded by air (lung)
The voltage of QRS and of ST segments fall proportionately.
Thus, the 0.5 mm of ST elevation in lead V9 (labelled V6) meets "criteria" for posterior MI.
You can also see that V3 still has much STD.
Had there been zero STE in V9, but persistent STD in V3, I would have called V9 a false negative!

Diagnosis: Posterior MI, right?

Just prior to transport, the patient became confused and agitated and, although blood pressure and pulse were OK, I was worried about cardiogenic shock.  We intubated him.

Cath lab
The BP was 70/40 on arrival to the cath lab and received a balloon pump and norepinephrine.


Left main: Severe calcific stenosis of ostial and distal left main
LCX: chronically occluded and filled by right to left and left to left collaterals
LAD: large caliber vessel with severe calcific stenosis of the proximal LAD with TIMI2 flow. There are large caliber diagonal branches with no significant stenosis. Mid LAD has another area of focal moderate stenosis. 
RCA: Moderate caliber with moderate to severe stenosis of the proximal to mild RCA. It gives epicardial collaterals to OM1 and OM2
All arteries had at least TIMI-2 flow, so this does not appear to be OMI to the posterior wall

Intervention: on Left Main and LAD

Next day

The highest troponin I recorded was 13.7 ng/mL.  

Normal LV size, normal LV wall thickness, mildly to moderately reduced LV
systolic function. Estimated EF 42%.
Wall motion abnormality - anterior, anteroseptal, apex
Wall motion abnormality - lateral (not posterior) 

I suspect that "lateral" means the lateral wall that has always been called lateral, and not the posterior wall that is now also called "lateral" on echo.  

In other words, I believe this ultrasound result refers to the lateral wall that is recorded by I, aVL, V5, V6; terminology has changed such that anything that is either posterior or lateral is now called "lateral."  This is unfortunate in my view, as it complicates the retrospective analysis of the ECG!!  Lateral can now include the wall that is detected by STE in I, aVL, V5, V6, but now also the wall detected by ST Depression in V2-V4)


It seems that in this case is an exception to the rule that if the ST depression is maximal in V2-V4 that it is a posterior MI.  Subendocardial ischemia does not localize, but that does not mean that subendocardial ischemia of the anterior wall (LAD) cannot manifest with STD in the anterior leads.  Only that it need not!

MY Comment by KEN GRAUER, MD (4/24/2020):
Prediction of a “culprit lesion” is an imperfect science. Striving to envision what prompt cardiac catheterization is likely to reveal in a patient with acute cardiac chest pain is enticing to the ECG enthusiast, and it does enhance clinical assessment and optimal management. But a more important goal is knowing when to (and when not to) activate the cath lab.
  • The cath lab was not activated in the prehospital setting in this case — despite a prehospital tracing identical to the initial ECG in the ED, that I have reproduced for clarity in Figure-1.

Figure-1: The initial ECG in this case (See text).

On seeing ECG #1 — Dr. Smith promptly activated the cath lab. THIS was the correct response — which was borne out a little bit later by sudden deterioration in the patient’s condition as efforts were being made to expedite transport to the cath lab.

MY THOUGHTS on ECG #1: I interpreted ECG #1 a bit differently than Dr. Smith. Clearly (as per Dr. Smith) — ST segment depression is maximal in leads V3 and V4 — and this finding should always suggest the possibility of acute posterior MI. This is especially true in ECG #1 because: i) the SHAPE of the depressed ST segment in the anterior leads here is consistent with a positive “Mirror Test” (Please see My Comment at the bottom of the February 16, 2019 post in Dr. Smith’s ECG Blog)andii) the R wave becomes relatively tall sooner-than-expected by lead V3 in ECG #1 (ie, potentially the mirror-image of an enlarging Q wave in the posterior lead distribution). That said:
  • I thought ST segment flattening and depression in ECG #1 was quite generalized — and quite marked in no less than 8 of the 12 leads (ie, ≥1-1.5mm ST depression in leads I, II, aVF, and V2-thru-V6).
  • The amount of ST elevation in lead aVR is marked (ie, ≥2 mm).
  • Instead of seeing at least some ST depression in lead V1 (as is typically seen with acute posterior MI) — there is slight ST elevation. This can occur, if what would have been ST depression in lead V1 is countered by ST elevation from associated acute RV involvement — yet there is no sign of acute inferior ST elevation (as I’d expect if there was acute proximal RCA occlusion causing RV infarction).
  • My Impression: While I thought acute posterior MI was a definite possibility — I favored the more nonspecific conclusion of diffuse subendocardial ischemia — which in this middle-aged man with increasingly severe chest pain, could be due to any of several possibilities: i) severe LMain disease; ii) severe proximal LAD disease; and/oriii) severe multi-vessel coronary disease.
  • BOTTOM Line: It really does not matter which of the above possibilities is the cause of the ECG findings we see in Figure-1. Regardless of the cause — the clinical situation cries out for prompt cardiac catheterization — which fortunately was expeditiously accomplished by Dr. Smith.

Recommended Resources