Monday, February 20, 2017

Chest Pain, LBBB, and a ratio that does not quite meet the Modified Sgarbossa Criteria


A 55 year old male heavy smoker presented with agonising chest pain and this ECG:
There is sinus rhythm.
There is left bundle branch block (LBBB)
There is no concordant ST elevation.
The highest ST/S ratio is in V3, at approximately 4.5/24 = 0.19
This is not a high ratio, but it is also not normal.

See this post:

A Fascinating Demonstration of ST/S Ratio in LBBB and Resolving LAD Ischemia

The mean maximal ST/S ratio in non-ischemic LBBB is about 0.11.  So 0.19 (19%) is abnormal.

In our validation study of the Modified Sgarbossa Criteria for diagnosis of acute coronary occlusion, we found that it performed similarly well using a cutoff of 20% or 25%:

Sensitivity and Spec at 20%: 84% and 94%
Sensitivity and Spec at 25%L 80% and 99%

So at a ratio of 0.19, there is still a high probability of occlusion.

I was sent this ECG by Facebook messenger, and asked my interpretation.  Here is my response:

"V3 is suspicious for excessive discordance. I would say it does not look like an acute STEMI, but I could be wrong."

The patient received thrombolytic therapy.

An angiogram was done after thrombolysis:

It showed moderate diffuse coronary disease and no thrombus.
LAD: mid segment moderate disease
OM: mid segment moderate lesion.
OM2: ostial moderate lesion
Ramus: moderate mid segment lesion
RCA: diffuse disease mid to distal, moderate lesion

The angiogram was considered to be "negative" for a culprit.

An ECG was repeated:
LBBB is resolved.
There is T-wave inversion in V2 and V3 highly suggestive of Wellens' waves
This represents LAD reperfusion

An echo was done:

There were septal, apical, and anterior wall motion abnormalities.

     --(high sensitivity troponin I BiomerieuxVIDAS TNHS):
99% reference = 19 ng/L (%CV = 7% at this level)
LoD = 2 ng/L

Initial: 13 ng/L (detectable, but still below the 99%)
Followup: 38 ng/L  (above the 99% and with a significant rise)

Here is the manufacturer's chart for interpretation:
You can see that a rise in high sensitivity troponin of greater than 10 ng/L is a "rule in."

Even with a "negative" angiogram, the weight of evidence heavily favors LAD occlusion at the time of the ECG:

1. There was excessive discordance, even if it did not meet the 20% or 25% cutoff
2. There was coronary disease, even if no thrombus; thrombus would likely be lysed by tPA
3. The followup ECG (most important) was consistent with reperfusion of the LAD
4. There were corresponding wall motion abnormalities.
5. The high sensitivity troponin had a diagnostic delta, even though the absolute level was minimally elevated

Learning points

1.  The cutpoints of 20% and 25% for the Modified Sgarbossa criteria maximize specificity, but are not fully sensitive for acute coronary occlusion.  Every case must be evaluated carefully.

2.  Not all ACS has a clearly visible culprit.

3.  The best way to assess whether ST elevation represents ischemia is to look at followup ECGs.  If the ST elevation resolves or evolves, then it is ischemic, even with all negative troponins!

See my last case:

Chest pain, ST elevation, and negative serial trops: normal variant ("early repol"). Right?

Friday, February 10, 2017

Chest pain, ST elevation, and negative serial trops: normal variant ("early repol"). Right?


A 58 year old man presented with intermittent chest pain for 2 weeks. He has active pain at the time of this initial ECG:
QTc is 455 ms
What do you think?


To me this is clearly an anterior STEMI, and it meets STEMI criteria even for someone under age 40 (at least 2.5 mm in V2 and V3, as measured at the J-point, relative to the PQ junction).

On the other hand, normal variant ST Elevation (often called early repolarization) may also have very marked ST elevation.   So when there is upward concavity in all of V5-V6, absence of any ST depression, and absence of Q-waves, it still might be early repol and the computer might not call this anterior STEMI.  Even the most contemporary algorithms are very inaccurate (see references below).

Thus, it is useful to use the STEMI-early repol calculator
(which is not, as far as I know, programmed into automated interpretation algorithms):

You can find the calculator here:
--- (
--- Or use the free iPhone app ("subtleSTEMI):
---Or go to www.mdcalc.com

The QTc = 455
ST Elevation at 60 ms after the J-point (STE60V3) = 4mm
R-wave amplitude in V4 (RAV4) = 17 mm

Formula value = 26.1.  At greater than 23.4, this is diagnostic of LAD occlusion until proven otherwise.

Clinical Course

The computer did not diagnose STEMI.  It did say "consider anterior injury."  However, the physicians thought it was early repolarization and admitted the patient for rule out MI.

Serial troponins were all undetectable (Beckman Coulter Access AccuTnI+3 Troponin I on DXL 600), LoD 0.010 ng/mL, 99% reference at 0.040 ng/mL), and thus the patient did rule out for MI.

At some point, the symptoms resolved; it is unclear when.

Fortunately, they recorded a second ECG 12 hours after the first:
Notice that all ST elevation has resolved!

The physicians were alarmed by this and realized they may have dodged a bullet.  They took the patient to the cath lab and found an 80% thrombotic LAD lesion.  It was stented.

This patient (and his physicians) were very lucky.  Had this patient not spontaneously reperfused, he would have lost his entire anterior wall, and possibly died.

It is dangerous to rely only on troponins for the diagnosis of acute coronary syndrome!

Learning Points:

1. Not all ischemic ST elevation results in elevated troponin
2. Unstable Angina still exists!!  Troponins may all be negative even with severe ACS.
3. Use the formula
4. Serial ECGs should be every 15 minutes, NOT every 12 hours!
5. High sensitivity troponins might have made a difference.  But maybe not.
6. Often, the only way to diagnose acute MI is with serial changes in the ECG.  In this case, resolution of ST elevation was diagnostic even in the absence of troponin elevation.


Contemporary computer algorithms are insensitive (65%) for STEMI, and only approximately 90% specific:

1. Mawri S, Michaels A, Gibbs J, et al. The Comparison of Physician to Computer Interpreted Electrocardiograms on ST-elevation Myocardial Infarction Door-to-balloon Times. Critical Pathways in Cardiology 2016;15:22-5.

2. Garvey JL, Zegre-Hemsey J, Gregg RE, Studnek JR. Electrocardiographic diagnosis of ST segment elevation myocardial infarction: An evaluation of three automated interpretation algorithms Journal of Electrocardiology 2016;49:728-32.

Sunday, February 5, 2017

Inferior MI? Or LVH?

A middle-aged male presented with acute chest pain:
There is sinus rhythm.
There is LVH by voltage in aVL, with what may appear to be secondary repolarization abnormalities
(so-called "LVH with strain", but this is a misnomer because it implies ischemia.  "LVH with secondary repolarization abnormalities" is more appropriate).
It has the "hockey stick" morphology typical of LVH with secondary repolarization abnormalities.

What do you think?

Here is the ECG of another middle-aged male with acute chest pain:

It is similar, but notice how the T-waves are not nearly as proportionately large as in the above ECG.

What do you think?

The bottom ECG is that patient's baseline LVH.

The top ECG is LVH with superimposed inferior acute MI.   The T-waves (both upright and negative ones) are far too large in proportion to the QRS.  There is also a concordant T-wave in lead II, and ischemic appearing biphasic T-waves in V4-V6, with a flattened ST segment in V3, all suggesting posterior and lateral involvement.

Technically, it is not "STEMI" because the ST elevation at the J-point is less than 1 mm.

This (top) ECG was missed by several interpreters and the patient had very delayed reperfusion therapy.  

The culprit was 100 % thrombotic occlusion of the mid RCA.  The peak troponin I was 47 ng/ml (very high).  There was a regional wall motion abnormality in the inferior and posterior walls.   The LAD was not involved.

The patient was discharged with a diagnosis of NonSTEMI.

There actually was a previous ECG for comparison, which proves the point.  Here it is:
Very different from LVH with superimposed inferior STEMI.

Here is the post-PCI ECG:
Notice inferior reperfusion (inverted) T-waves
Notice precordial large T-waves (posterior reperfusion T-waves)
Notice lateral reperfusion T-waves.
These are "Wellens' waves" of inferior, posterior, and lateral walls.

Learning point

Know the T-wave to QRS proportions in LVH vs. LVH with superimposed MI.

Thursday, February 2, 2017

Patient with severe DKA, look at the ECG

This patient presented with severe DKA.

Here is the ECG:
Sinus tachycardia.
What do you think?

The computer and physician reader wrote: "ST depression, consider subendocardial injury."
The computer read the QT as 365 ms and the QTc as 424 ms.  What else?

I read the QT interval as somewhere between 480 and 580 ms, depending on the complex, with a QTc (Bazett correction) of 630 - 763 ms.  There is a very prominent U-wave and some of what may appear to be a QT interval is a QU interval.  So the real QT is shorter, but the computer does not mention the U-wave, and the U-wave is as important as the T-wave in predicting cardiac dysrhythmias.

This is an extremely dangerous ECG.

The K returned at 1.9 mEq/L.

This is extremely low for DKA.

K in DKA is usually high from shifting out of cells, and will go lower as it shifts into cells during treatment.

Therefore, hypokalemia in the setting of DKA is truly life threatening and must be treated aggressively.

When the ECG shows the effects of hypokalemia, it is particularly dangerous.  In spite of aggressive K replacement, the patient went into ventricular fibrillation.


See this post: STEMI with Life-Threatening Hypokalemia and Incessant Torsades de Pointes

I could find very little literature on the treatment of severe life-threatening hypokalemia.  There is particularly little on how to treat when the K is less than 2.0, and/or in the presence of acute MI.

Here are the American Heart Association Guidelines: 

Part 10.1: Life-Threatening Electrolyte Abnormalities

Treatment of Hypokalemia

"The treatment of hypokalemia consists of minimizing further potassium loss and providing potassium replacement.  IV administration of potassium is indicated when arrhythmias are present or hypokalemia is severe (potassium level of less than 2.5 mEq/L).  Gradual correction of hypokalemia is preferable to rapid correction unless the patient is clinically unstable.

"Administration of potassium may be empirical in emergent conditions.  When indicated, the maximum amount of IV potassium replacement should be 10 to 20 mEq/h with continuous ECG monitoring during infusion  A more concentrated solution of potassium may be infused if a central line is used, but the tip of the catheter used for the infusion should not extend into the right atrium.

"If cardiac arrest from hypokalemia is imminent (i.e., malignant ventricular arrhythmias are present), rapid replacement of potassium is required.  Give an initial infusion of 10 mEq IV over 5 minutes; repeat once if needed.  Document in the patient's chart that rapid infusion is intentional in response to life-threatening hypokalemia."

CASE: Prehospital Cardiac Arrest due to Hypokalemia

I recently had a case of prehospital cardiac arrest that turned out to be due to hypokalemia.
We could not resuscitate her, but we did have excellent perfusion with LUCAS CPR, such that pulse oximetry had excellent waveform and 100% saturations, end tidal CO2 was 35, and cerebral perfusion monitoring was near normal throughout the attempted resuscitation.

During the resuscitation, I ordered 10 mEq KCl push, but the patient received 40 mEq of KCl, push (far more than recommended)  The resident had ordered 40 mEq and that is what the nurses heard.

Is 40 mEq too much? Or the right amount?

Contrary to my expectations, after pushing 40 mEq, the K only went up to 4.2 mEq/L.

What is the right amount of K to push in life-threatening hypoK?
In a 70 kg person, there are 5 liters of blood and 3 liters of serum.  Since it takes some time (how long?) for K to shift out of the intravascular space into the interstitial space and then into the intracellular space, 3.0 mEq of K pushed fast and circulated theoretically would raise serum K immediately by 1.0 mEq/L, and 10 mEq would increase it by 3.3 mEq/L, from 1.9 to 5.2.   Thus, 40 mEq should raise it by 13 mEq/L!! 

But this is before redistribution to the interstitial space.

As I indicated above, in our cardiac arrest case, after pushing 40 mEq, the K only went up to 4.2 mEq/L.   
There are about 13 liters of extracellular fluid in a 70 kg person (10 liters interstitial fluid + 3 liters serum).  So if K redistributes very quickly to this extracellular space, then 40 mEq is appropriate.

The difficulty is in estimating the ongoing shift.  As you infuse K, it will start to shift into depleted cells and the serum K will fall again rapidly.  Thus, it is critical in patients like this to repeatedly and rapidly, after each bolus, measure the K, and supplement as needed.

Total Body Potassium: a 70 kg person has about 7500 mEq of total body K, but the extracellular fluid has only about 45 mEq!   Of course the difficulty with K replenishment is that the total body stores may be depleted by far more than can possibly be quickly repleted.  The estimated deficit associated with a serum decrease from 4.0 mEq/L to 3.0 mEq/L is 100-200 mEq of total body K, and from 3.0 mEq/L to 2.0 mEq/L, the associated loss is double, at 200-400 mEq.* [Sterns RH, et al. Internal potassium balance and the control of the plasma potassium concentration. Medicine (Baltimore) 1981;60:339-54].  

But 100 mEq given all at once by bolus, could (before any redistribution to interstitial space) raise the serum K by 33 mEq/L (and be immediately fatal)!!

*The NEJM review referenced below (and ACLS, for what that is worth), states that, on average, in a "typical" 70 kg person, the serum K falls by 0.3 mEq/L for every 100 mEq total body deficit.  However, this review references the Sterns article above, which by my reading does not state this.

Further complicating the issue is that severe hypokalemia can result in rhabdomyolysis and subsequent K release, with resulting hyperkalemia!

Here is review of hypokalemia from the NEJM, but it is mostly about etiology, and says little about rapid replacement in life-threatening hypokalemia EXCEPT to emphasize how dangerous rapid replacement is.

Friday, January 27, 2017

Grouped Beating, What is it?

This is a case seen by one of my great partners at Hennepin County Medical Center Dept. of Emergency Medicine, Dr. Ashley Strobel, @AStrobelMD.


A patient was found down and was quite ill and in shock.  The POCUS of his heart showed very poor contractility.

Here is his initial ECG:
It is a supraventricular rhythm, with grouped beating.
What else?

Here is a rhythm strip (10 seconds of all leads):
Again, there is grouped beating.
What else do you see?

When she showed me this ECG, I was at first distracted by the rhythm, and did not immediately see the life threatening finding.  It took me a minute before I looked beyond the rhythm.

What is it?

The QRS is slightly wide, greater than 120 ms.  Leads V1 and V2 have domed ST Elevation with T-wave inversion that is very similar to Brugada pattern ("Brugada phenocopy"), similar to my last post.

Brugada pattern should make you think of either hyperkalemia or sodium channel blockade.  Additionally, there is T-wave peaking in many leads.

Dr. Strobel suspected hyperkalemia and treated with Calcium and shifting. She was right.  The K returned at 7.1 mEq/L.

Here is another post with many cases of Brugada phenocopy due to hyperkalemia:

This ECG is NOT Pathognomonic of Brugada Syndrome

Here are other etiologies of Brugada phenocopy:

Analysis of rhythm

Here is Ken Grauer's analysis:

"While there most definitely is a repetitive pattern of group beating (which is a feature of Wenckebach) — this is NOT completely typical of Wenckebach conduction — because the pause containing the dropped beat is NOT less than twice the shortest R-R interval. That’s not to say there can’t be some component of Wenckebach conduction out of the AV node — but rather to recognize that lack of this above “footprint” is indeed a bit atypical …. and under normal circumstances is often a clue that the mechanism may not be Wenckebach despite group beating ...

"That said — T waves in many leads are tall, peaked and pointed with symmetric upstroke and downstroke and narrow base in a number of leads in this somewhat widened QRS rhythm with group beating but no P waves. This strongly suggests hyperkalemia (which may also be responsible for the Brugada-like V1,V2 findings as well).

"It’s been my experience that most of the time it is NOT worth one’s while to contemplate arrhythmia mechanisms when the underlying problem is hyperkalemia — because this electrolyte abnormality “does not follow the rules” — and because whatever arrhythmia abnormality we see will “go away” once you fix this underlying problem of hyperkalemia. Sounds like this is precisely what happened in your case."

Learning Point:

Always be on the lookout for hyperkalemia.  It comes in many form and can mimic many pathologies.

Wednesday, January 25, 2017

A Patient with Cocaine Chest Pain and Prehospital Computer interpretation of ***STEMI***

A 20-something male drank heavily of ethanol and used cocaine, then was involved in a stressful verbal altercation, at which time he developed chest pain.

911 was called and the medics recorded this ECG (unfortunately, leads V4-V6 are missing)
Due to marked ST Elevation, the computer read was ***STEMI***
What do you think?

He arrived in the ED and had this ECG recorded:
Very similar to the prehospital ECG.
The Mortara (Veritas algorithm) Interpretation was:

 ***ACUTE MI***
What do you think?

The ECG shows Brugada morphology in V1 and V2, and the typical normal variant ST elevation in lead V3.

Brugada morphology can be caused by baseline Brugada morphology, including Brugada syndrome, or by hyperkalemia or Sodium channel blockade.

Cocaine not only has effects on dopamine neurotransmission, but is also a sodium channel blocker, as are all "-caine" local anesthetics.  Cocaine is well known to result in Brugada morphology.

See this post and associated case reports:

Cardiac arrest, severe acidosis, and a bizarre ECG

The patient was admitted and ruled out for acute MI by serial troponins.

Below are subsequent ECGs, showing resolution of the Brugada morphology as the cocaine metabolizes.  Cocaine metabolism is rapid.  After approximately 3-4 hours, the cocaine and its effect are gone.  Testing for cocaine is for the inactive metabolite Benzoylecgonine, and this inactive metabolite is present for days.  So a positive screening test for "cocaine" does not imply persistent intoxication.

Here are the serial ECGs:
Time 1 hour:
Cocaine Brugada Effect is still present

Time 4 hours:
Minimal effect is still present

Time 10 hours:
The ECG only shows some slight abnormalities in V1 and V2, with minimal residual saddleback morphology in lead V2.

The vast majority of cocaine chest pain is NOT due to myocardial ischemia or infarction, and most is in young males with normal variant ST Elevation or LVH, so it often looks scary on the ECG.

As there was no personal history of syncope or family history of sudden death, the patient was discharged with cardiology followup.

Sunday, January 22, 2017

A very fast narrow complex tachycardia in an Infant

Case 1

A 4 month old infant with known co-arctation of the aorta and reflux presented with respiratory distress.

Here is the ECG:
A Narrow Complex Tachycardia with a rate of 218.
What is the rhythm?

Answer: there are clear P-waves in nearly every lead.  Notice the intervals are very short, which is typical of infants.  Infants can have very fast sinus tachycardia, easily reaching a rate of 220.  Most SVT in infants is faster than 220.

Throughout this case, the patient was on a cardiac monitor and the rate drifted up as high as 246, still in sinus (not recorded)!

The tachycardia turned out to be a result of disease (pneumonia), not a cause of it.

Case 2

This newborn presented to the ED for some exudate on the umbilical stump. There was no fever.  The infant was very well appearing.  The palpated heart rate at triage was 140.  There were no respiratory symptoms.

The ECG was performed because, on auscultation, the heart rate was far higher than palpated by pulse.
Narrow complex tachycardia at a rate of 300.

This shows how fast SVT can go in an infant and how well tolerated it might be.  This infant was completely without any signs of illness.

There is also electrical alternans, which is a normal finding in PSVT.

If you see this in sinus tachycardia, it indicates tamponade.

It was converted to sinus with adenosine.

Learning points

This shows how well a newborn/young infant tolerates a very fast heart rate, and that sinus tachycardia of 218 is not unusual, and may even go as high as 240!

If in doubt, use Lewis leads to find otherwise hidden P-waves

I don't have a case of use of Lewis leads in pediatrics, but here is a nice one in an adult:

Wide Complex Tachycardia. What is the Diagnosis? Use of the Lewis Lead.

Friday, January 20, 2017

Do patients with LBBB and STEMI, when reperfused, develop T-wave inversion (reperfusion T-waves)?

The case below was contributed by Pendell Meyers, an EM G1 at Mt. Sinai (the case did not come from Mt. Sinai though!)
Pendell is the lead author on our Modified Sgarbossa Criteria Validation Study.

Meyers HP.  Limkakeng AT.  Jaffa EJ.  Patel A. Theiling BJ. Rezaie SR. Stewart T. Zhuang C.  Pera VK.  Smith SW.  Validation of the Modified Sgarbossa Rule for Diagnosis of STEMI in the Presence of Left Bundle Branch Block. American Heart Journal 170(6):1255-1264; December 2015.

Before the case, a few comments:

Pendell and I just published a case report of a patient with left bundle branch block who presented with chest pain that then resolved.  His ED ECG showed his baseline LBBB, with no evidence of MI.  Over the ensuing hours, he developed classic T-wave inversion of Wellens' syndrome, but in the context of LBBB!  Troponins were then positive, and the angiogram revealed a 99% LAD lesion with thrombus.  The case demonstrates that Wellens' syndrome can occur in the context of LBBB.

Here is a link to the case report:

Dynamic T-wave inversions in the setting of left bundle branch block

Though Wellens' syndrome was described in the LAD territory, I have shown cases demonstrating that it occurs in any coronary distribution.  That is to say, that reperfusion results in terminal T-wave inversion even if the involved territory is the inferior or the lateral wall.

Today's case shows how, even in the context of LBBB, reperfusion of the inferior and lateral walls, just as with the anterior wall, can result in typical reperfusion T-waves (which is what Wellens' waves represent).

Wellens' syndrome represents a post (spontaneous) reperfusion state.  This is why the patient is always pain free.  The ST elevation is implied but was never recorded because the patient did not have an ECG during pain.

This is worth a look (7 serial ECGs showing evolution of Wellens' over time):

Classic Evolution of Wellens' T-waves over 26 hours


An 82 year old had the acute onset of chest pain.  Here is his first ECG, time zero:
There is concordant ST elevation in leads II, V5 and V6.  It may not reach a full millimeter, but the QRS is so small that we should make an exception here.
It is proportionally large concordant ST elevation!

The cath lab should have been activated at this point, but apparently it was not.  Instead, another ECG was recorded at time 46 minutes:
Now there is more than 1 mm of concordant ST elevation.
In addition, there is now excessively proportionally discordant (more than 25% of preceding S-wave) in leads III, and aVF.
So there is a definite inferior and lateral MI.  There is no ST depression in lead I, which suggests a circumflex lesion

The cath lab was activated and a circumflex occlusion was opened and stented, with a door to balloon time of 3 hours.

Here is the first ECG recorded after reperfusion:
ST deviation has resolved. 
There is T-wave inversion in III and aVF, typical of "Inferior Wellens'"
There is no T-wave inversion yet in the lateral leads.
There is increased T-wave amplitude in V2 and V3: these represent posterior reperfusion T-waves, or Wellens' syndrome of the posterior wall.

2 Examples of Posterior Reperfusion T-waves

(these are in the context of normal conduction, not LBBB)

Here is our formal study of posterior reperfusion T-waves, written by me and lead author Brian E. Driver (Hennepin) and published in Emergency Medicine Journal:

Posterior reperfusion T-waves: Wellens' syndrome of the posterior wall

Here is a still later ECG after reperfusion:
There is evolution of T-wave inversion (deeper) in the affected leads.
This is analogous "Wellens' waves" of the inferior and lateral leads, in the presence of LBBB!

Peak troponin T was 3.71 ng/mL, which is indicative of a large MI (Troponin T peaks are far lower than troponin I).  Earlier recognition of the concordant ST elevation could have saved more myocardium.

With reperfusion, T-wave inversion even occurs in LBBB:

As an explorative substudy of our validation of the modified Sgarbossa criteria, Pendell and I and others studied T-wave inversion.  We looked at serial ECGs on patients with acute coronary occlusion ACO) who underwent reperfusion and compared to serial ECG on patients without ACO.  Unfortunately, as a result of our multisite study in which ACO came from many institutions and controls from one institution, only 6 of 45 patients with ACO and reperfusion had serial ECGs available, and all 245 patients without ACO had serial ECGs available.

When this pattern was retrospectively defined as being either 1) present in at least two
contiguous anterior or inferior leads in at least two consecutive ECGs prior to reversal or 2) deeper than 3 mm in two contiguous leads (requiring only one ECG), it was found to be predictive of reperfused ACO (either spontaneously prior to catheterization or with mechanical reperfusion) with derived sensitivity and specificity of 5 of 6 [83% (95% CI 36–99%)] and 241 of 245 [98% (95% CI 96–99%)]. 

Meyers HP et al. Evaluation of T-Wave Morphology in Patients With Left Bundle Branch Block and Suspected Acute Coronary Syndrome

Another very illustrative case:
Here is a case of transient LAD occlusion that resulted in transient LBBB.  After reperfusion, Wellens' waves are evident:

Chest pain and LBBB. LBBB resolves and there is V1-V3 T-wave inversion.

Recommended Resources