Chest Pain and a Very Abnormal ECG

A male in his 60's with no history of coronary disease presented with chest pain.  Here is his ED ECG:
He had several previous ECGs this year and they were all identical.

What do you notice?  What should you do?

There is bizarre ST depression, T-wave inversion, prolonged QT, and large U-waves (vs. VERY long QT).

This should make you worry about hypokalemia or about acquired long QT syndrome.  I saw this ECG in a stack and noticed that caregivers had been only worried about ischemia.  The patient was appropriately evaluated for ischemia (with a negative workup) but other etiologies of this ST depression had not been considered, and the long QT and U-waves had not been noticed.  This is partly because the computer interpretation did not notice the U-waves.  The computer did measure the QT accurately as 488 ms.  However, it should EITHER have noticed the U-waves, OR included them in the QT interval. If it had, the QTc would have been over 600 ms.

Seeing this, I thought it would be hypokalemia, but the K was normal.  So I looked at the medications and found that the patient was on methadone at a dose of 260 mg daily.

Doses of 60 mg daily are associated with a longer QT and with U-waves.

Methadone may cause torsade de pointe due to these effects.  See this case of Polymorphic VT with long QT (Torsade) due to Methadone.

The patient's physicians were contacted and he was taken off methadone.

Learning points:
1. U-waves are not only caused by hypokalemia, but also by drugs that block postassium channels.
2. The computer does not accurately measure QT and QU intervals.
3. Just because an ECG findings is stable, seen on old ECGs, does not mean it is safe.

Some Cardiologists still are not familiar with Sgarbossa Criteria…..

This case shows how you have to advocate for your patient. You have to be the expert.  I talk to a lot of emergency physicians who have trouble convincing their interventionalist colleagues of various ECG findings.  

Many cardiologists and even interventionalists, are not aware of many of the newer findings of STEMI equivalent, including the Sgarbossa criteria, and especially are unaware of the modified Sgarbossa criteria.  

My partners and I are fortunate that at our institution there is a long history of ECG expertise in our cardiology department.  We also have a very cooperative system in which all concerns of the emergency physicians are addressed.

It is very important to establish these relationships so that when difficult cases arise, the patient gets the best care.

Case presentation

A male in his mid 40's presented with 20 minutes of chest pain.  He had no previous cardiac disease.  Here is his first ECG:
There is sinus rhythm with LBBB.  There is subtle concordant STE in lead aVF, almost 1 mm.  There are about 2 mm of discordant STE in III.  Since the S-wave is only 3 mm, this is proportionally excessively discordant ST elevation.

One may not be convinced of STEMI with less than 1 mm of concordant STE and 2 mm of discordant STE, but it should certainly catch your attention.

Within minutes, the patient had a ventricular fibrillation arrest and was resuscitated.  This is the post-resusitation ECG:

There is sinus tach and LBBB still.  Now there is huge concordant ST elevation in III, also in aVF, and huge reciprocal concordant ST depression in aVL.  There is concordant STE in V3 and excessively proportionally discordant ST depression (greater than 30% ratio) in V5 and V6.

This second ECG is diagnostic of STEMI.

The physician activated the cath lab in the middle of the night.

When the interventionalist heard that the patient had LBBB, he was furious and stated that you cannot diagnose STEMI in the presence of LBBB (even though there was a cardiac arrest).

When he arrived and saw the ECG, he insisted that all ECG findings were due to post-resuscitation changes.  (Aside: while it is true that cardiac arrest can cause strange ST elevation, one can easily differentiate it from actual coronary occlusion by obtaining serial ECGs.  If due to cardiac arrest only, it should quickly start to normalize).  

He insisted that "You inappropriately called in the cath team.  The only STEMI by criteria is 2mm ST elevation in 2 consecutive leads with normal QRS AND it is the cardiac arrest that created these ECG abnormalities." 

The emergency physician insisted that it was STEMI and instructed on the Sgarbossa criteria.  

He did take the patient to the cath lab and found a 100% acute RCA occlusion.

He still insisted that there are no ECG findings in of STEMI in patients with LBBB except for New LBBB.

She gave him some papers to read.  We now know that new LBBB has very weak correlation with acute occlusion, and that + Sgarbossa criteria is very specific for occlusion, and that the modified Sgarbossa criteria are much more sensitive.

Here is a calculator from

Learning point: You need to be the ECG expert and advocate for your patient
The emergency physician's confidence in her diagnosis could have been undermined by the interventionalist.  She could have doubted herself and relented.  But she had learned about LBBB and Sgarbossa criteria well enough to be confident.  She also was aware that consultants are fallible, and have various occasional weaknesses.  So she was able to be persistent in advocating for the patient, and to do so diplomatically, so that the interventionalist would not resist.

Imagine how resistant the interventionalist would have been if the patient had no cardiac arrest, but rather chest pain only.

Is it Wellens’ Syndrome?

A middle-aged African American man with history of tobacco use, HTN, and chronic renal disease, but no known coronary disease, presented with chest pain. 

Symptoms in the 24 hours prior to presentation: On the evening prior to admission, he had an episode of sharp stabbing chest pain while talking to his wife, which improved after 30 minutes. This recurred on the morning of the date of presentation.  He did not seek care for these episodes of pain.

On the evening of presentation, the chest pain recurred at work, with increased pain radiating to his shoulder and neck in the afternoon, associated with diaphoresis and SOB.  He was given ASA and nitro by medics with improvement.  He was pain free on arrival.

Here is his initial ED ECG:
1. Computerized QTc is 391.  There is a bit of ST elevation in V2 and V3, and the T-wave turns sharply downward, with a trace of terminal T-wave inversion in V3.

I was slightly concerned for the subtle beginnings of Wellens' waves in V2 and V3, so recorded this one 37 minutes later:

2. Now the QTc is 411 and there is definite terminal T-wave inversion in V3, and some also in V4 and V5.  Importantly, there is none in V2.

At this point, I was worried that this was "Wellen's syndrome" and initiated medical treatment for ACS.

Wellens' is a syndrome and not an ECG finding.  If you only look at the ECG, there will be many false positives.  In order for the diagnosis to be fairly specific, there must be:

1. Chest pain (not some other atypical symptom) of at least 20 minutes duration
2. Complete resolution of pain at the time of the ECG
3. Presence of normal R-waves in the precordial leads (not Q-waves, not LVH)
4. Pattern A or Pattern B (see bottom of post) T-wave inversions in V2-V3, and also possibly, in decreasing order of prevalence, V3, V4, and V5.
5. Evolution of the pattern [though de Zwaan and Wellens' did not describe evolution (probably because they did not do serial ECGs nearly as often as we do today), in my experience Wellens' always evolves: biphasic T-waves (Pattern A) become deep symmetric (Pattern B) and they also extend out to V4, then V5, then V6.]

The patient remained pain free, and this was recorded 27 minutes after the 2nd ECG (64 after the first):

3. QTc is 402.  The terminal T-wave inversion is gone

At this point, without any more evolution, I considered that this might not be Wellens' syndrome, which in a persistently pain free patient usually continues to evolve, and not resolve.

I thought I might have been deceived by a Wellen's mimic.

So I decided to look back at ECG 2 and realized that I had, indeed, been deceived.

What is the real diagnosis?

This is typical benign T-wave inversion.  It is impressive that it can change from one ECG to the next.  This BTWI was dynamic: in some ECGs, there is terminal T-wave inversion and in others, there is not

Left Ventricular Hypertrophy is the other common mimic of Wellens' syndrome.  Here is a representative case.

Here is another case of benign T-wave inversion, with a detailed discussion.

In both ECGs, V3 and V4 have the classic look of BTWI, which particularly involves very tall T-waves, and almost no S-wave, in the leads with biphasic T-waves.

Beware, however, because unstable angina can also have dynamic t-wave inversion that mimics Wellens' waves, which appear and disappear.  See this case.

For the classic evolution of Wellens' waves over 26 hours, see this case.

Case continued

This 4th ECG was recorded 46 minutes after the 3rd ECG (110 minutes after the first):
4. No major difference

The patient remained asymptomatic, but his ECG is dynamic:
5.  Recorded 4 hours after arrival, and 2 hours after ECG #4.

He ruled out (all contemporary troponins less than 0.03 ng/mL).  In spite of this, because of his ECG, he underwent an angiogram.  It was completely normal.

Here is one week later:
6. Now there is no trace of terminal T-wave inversion!!

Typical Wellens'

Here is classic Wellens', with evolution:
Initial ECG, and two subsequent ECGs hours apart in True Wellens'.  Notice the evolution extending farther out to more lateral precordial leads, and also from biphasic (Wellens' Pattern A) to deep symmetric (Wellens' Pattern B).  [Later writers mistakenly gave a label of Type 1 to Pattern B and Type 2 to Pattern A]

Here is Wellens' original image:

Here is another post on Benign T-wave inversion: Wellens' syndrome - NOT!

Learning points:
1. Not all that looks like Wellens' is Wellens'
2. Wellens' is a syndrome, not an ECG finding.
3. If it is truly Wellens,' it will evolve and there will almost always be postive troponins.
4. Even if the whole syndrome is present, as in this case, there are false positives. Benign T-wave Inversion (BTWI) is one.
5. BTWI can be dynamic.
6. Unstable angina (troponins all negative) can also have dynamic T-wave inversions that appear and disappear
7. There are identifying features of BTWI.  I list them below:

BTWI is a normal variant associated with early repolarization.  K. Wang recently studied it.  He reviewed ECGs from all 11,424 patients who had at least one recorded during 2007 at Hennepin County Medical Center (where I work) and set aside the 101 cases of benign T-wave inversion.  97 were black.  3.7% of black men and  1% of black women had this finding.  1 of 5099 white patients had it.  Aside from an 8.8% incidence (9 of 109) black males aged 17-19, it was evenly distributed by age group.

I have reviewed these 101 ECGs, and what strikes me is:

1. There is a relatively short QT interval (QTc < 425ms)  
2. The leads with T-wave inversion often have very distinct J-waves.
3. The T-wave inversion is usually in leads V3-V6 (in contrast to Wellens' syndrome, in which they are V2-V4)
4. The T-wave inversion does not evolve and is generally stable over time (in contrast to Wellens', which always evolves). 
5. The leads with T-wave inversion (left precordial) usually have some ST elevation 
6. Right precordial leads often have ST elevation typical of classic early repolarization
7. The T-wave inversion in leads V4-V6 is preceded by minimal S-waves
8. The T-wave inversion in leads V4-V6 is preceded by high R-wave amplitude
9. II, III, and aVF also frequently have T-wave inversion. 

A Mystery Rhythm, Explained by K. Wang’s Ladder Diagram.

This was presented at the SMACC-Gold EKG workshop:
Can you figure it out?
See the answer below.

Here is a nice ladder diagram and explanation from K. Wang:

--The primary problem is irregular sinus bradycardia (sinus node dysfunction), with junctional escape beats (R1, 2, 3, 5, 6, 7,and 9, which occur with exactly the same interval) and two capture (conducted) beats (R4 and 8). 
--The patient also must have some degree of AV conduction problem both anterograde and retrograde. Otherwise, R2, 3, 6, and 7 would have resulted in retrograde P waves, or P2 and 5 would have conducted. 
--P3 and 6 occur with slightly longer RP interval than P2 or 5, which is the reason why they are conducted but still with a long PR interval.
--P1, 4, and 7 are completely assumed even though it is unusual for them to occur exactly within the QRS (too much of coincidence). Otherwise, P3P5 interval is too long not to have a P wave in between when P2P3 or P5P6 can occur.

Diffuse ST Elevation. What do you think?

What do you think of this ECG, without any clinical information?

There is inferior and lateral ST elevation and ST depression in V1 and V2, just what would be present in an infero-postero-lateral STEMI.  What other finding is there?


An extremely long QTc is present, and this was even noted by the computer algorithm, which accurately measured it at 575 ms.  This strongly suggests stress cardiomyopathy (SCM).  This immediately caught my eye when shown to me and SCM immediately came to mind.

The patient presentation is of course critical to the diagnosis of SCM: this patient presented with altered mental status and a Na of 107 mEq/L.  She had presumably had a seizure.

An echocardiogram revealed multiple wall motion abnormalities, though not the typical "Takotsubo" Apical Ballooning.

When I saw this ECG, I also thought "hypokalemia,"  Why?  There is a "wavy" pattern in V2 which is caused by an inverted T-wave followed by a large U-wave.  See this annotated version: 

The lines show the "wavy" pattern and the arrow points to the large U-wave (arrow).  The wavy pattern could be mistaken for a down-up T-wave, which is common in Posterior MI, but in this case it would be a far too long QT interval for this to be T-wave..

Here are some more examples of the wavy pattern of hypokalemia.

Peak troponin I was 2.5 ng/mL.  ECG findings quickly resolved.  There was no angiogram.

How to suspect takotsubo stress cardiomyopathy from the presenting symptoms, signs and ECG

First, it is believed to by caused by diffuse small vessel ischemia due to catecholamines, and thus has the same electrophysiologic substrate as STEMI.  It affects the entire heart except the base, resulting in diffuse circumferential wall motion abnormalities that only spare the base (top) of the heart.  These diffuse wall motion abnormalities (lateral, posterior, inferior, septal, anterior, apical) result in systolic "apical ballooning" which looks like a Japanese octopus trap (a takotsubo).

Of course, the clinical presentation can help to suspect this:
It is most common in postmenopausal women and is associated with severe emotional upset (stress) or severe physiologic stressors such as respiratory failure. 

Stress Cardiomyopathy (Apical Ballooning syndrome, or Takotsubo cardiomyopathy) only presents with ST elevation in about 1/3 of cases, but when it does, it is one of the most difficult mimics of anterior STEMI, and often the only way to tell the difference is to do an angiogram. 

ECG Differentiation of SCM from STEMI

First, a very long QT is quite specific for SCM, but not sensitive.  In this study, the mean QTc for SCM was 567 (+/- 81) ms vs. 489 (+/- 61) ms for anterior STEMI.

A review and analysis of the literature in ECG differentiation of the two entities (1-6) reveals that an analysis of ST elevation vector is a good predictor of anterior STEMI vs. stress cardiomyopathy.  In anterior STEMI, whether proximal, mid, or distal LAD occlusion, there is ST elevation in V1 (1 mm measured at 80 ms after the J-point) in about 80% of cases, whereas this is present in 20% of cases of SCM.  SCM also has a negative ST segment in aVR and is more likely to have ST elevation in inferior leads, or at least absence of ST depression in inferior leads (however, 40% of anterior STEMI lack inferior ST depression). Finally, precordial ST elevation in SCM is more pronounced in V3-V5 vs. V2-V4.

Putting these all together, it is apparent that the ST vector in anterior STEMI is more commonly anterior and superior (V1-V4), without STE in inferior leads, whereas in SCM the ST Vector it is more inferior and lateral (V2-V5, with STE in inferior leads and ST depression in aVR).  This correlates with the location of wall motion abnormalities in SCM (diffuse and toward the apex, similar to pericarditis, including inferior and lateral walls) vs. anterior STEMI (anterior as well as septal-apical). 

In this very interesting study by Kosuge et al., the combination of the presence of ST-segment depression in lead aVR and the absence of ST-segment elevation in lead V1 identified SCM (vs. anterior STEMI) with 91% sensitivity, 96% specificity, and 95% predictive accuracy, which was superior to any other electrocardiographic findings.

1.     Tamura A et al.  A New Electrocardiographic Criterion to Differentiate Between Takotsubo Cardiomyopathy and Anterior Wall ST-Segment Elevation Acute Myocardial Infarction.  Am J Cardiol Sept 2011; 108(5):630-633.

2.    Ogura R, et al. Specific findings of the standard 12-lead ECG in patients with “takotsubo” cardiomyopathy: comparison with the findings of acute anterior myocardial infarction. Circ J 2003;67:687– 690.

3.    Inoue M, et al. Differentiation between patients with takotsubo cardiomyopathy and those with
anterior acute myocardial infarction. Circ J 2005;69:89 –94.

4. Bybee KA, et al.  Electrocardiography cannot reliably differentiate transient left ventricular
apical ballooning syndrome from anterior ST-segment elevation myocardial infarction. J Electrocardiol 2007;40:38.e1–38.e6.

5. Jim MH, et al. A new ECG criterion to identify takotsubo cardiomyopathy from anterior
myocardial infarction: role of inferior leads. Heart Vessels 2009;24:124–130.

6. Kosuge M, et al.   Simple and accurate electrocardiographic criteria to differentiate takotsubo
cardiomyopathy from anterior acute myocardial infarction.  J Am Coll Cardiol 2010;55:2514 –2516.

Prolonged (63 minutes) Ventricular Fibrillation, Followed by Unusual Cardiogenic Shock

A middle-aged patient presented in continued ventricular fibrillation after 5 minutes of down time and 45 minutes of prehospital resuscitation by medics, using King Airway, LUCAS, Inspiratory Threshold Device (ITD, ResQPod), defibrillation 4 times, and epinephrine x 3 through an intraosseous line.  The patient had continued to swallow and breathe while being resuscitated (suggesting effective chest compressions).

In the ED, ventilations were found to be effective with the King airway.  LUCAS compressions were continued, and end-tidal CO2 was 26 mm Hg (supporting evidence that chest compressions were effective and supporting the possibility of a good neurologic prognosis). 

Ventilations were administered slowly at 10 per minute, following the indicator light on the ITD.  After 300 mg of amiodarone, 100 mg of lidocaine, and 500 mcg/kg of esmolol + 50 mcg/kg/min drip, plus more epinephrine and also bicarbonate, and another defibrillation, a rhythm check at 18 minutes after arrival revealed an organized, mostly narrow complex, but slow, rhythm.  No pulse could be felt.  A bedside cardiac ultrasound, subcostal view, showed the following:

This is reverse orientation: ventricles are right upper and atria left lower.  The RV is closest to the transducer and is very small (essentially excluding pulmonary embolism as etiology).  The LV has extremely thick walls and a very small LV chamber (there is very little blood to pump)

The emergency physician interpreted this as hypertrophic cardiomyopathy (HOCM), and was suspicious of HOCM as the etiology of arrest, as it is well known to cause ventricular fibrillation.

The vascular transducer was used to visualize the carotid artery, with power Doppler, and this showed good flow in the carotid, corresponding to the cardiac contractions.

An ECG was recorded:
The first 5 seconds appears to be an irregular wide complex tachycardia, with multiform QRS.  However, there are many beats which are clearly narrow complex but only appear to be wide due to ST segment shifts.  This is best seen in lead II across the bottom.

The next 5 seconds appear to be an irregularly irregular, polymorphic wide complex tachycardia.  This is reminiscent of Atrial Fib with WPW, except that the rate is not as fast as is usually seen with that entity.

The exact rhythm here is uncertain.

It is not unusual to have very bizarre ECGs immediately after resuscitation, especially if there is underlying cardiomyopathy, as suspected here.
Case continued:

The patient remained very hypotensive.  Due to the very small LV volume and the need for volume loading in patients with very thick-walled ventricles and slit-like LV [as one sees in HOCM (See this very instructive case)], volume loading was begun.

Another ECG was recorded 5 minutes after the first:
There is an uncertain supraventricular rhythm, sinus vs. accelerated junctional, with a narrow QRS.  There appear to be delta waves.  There is bizarre ST elevation in V1-V3 and aVR that does not look like STEMI
The emergency physician assessed that this was not a STEMI and did not activate the cath lab.  

As the patient was bradycardic and hypotensive, the patient received push dose epinephrine, and a norepinephrine drip was started.   The esmolol was stopped.  Fluids were continued.  The BP and pulse rose.

The venous pH was 7.16, pCO2 48, bicarb 17, and lactate 6.8.  K = 3.6 mEq/L.  

The King airway was removed and he was endotracheally intubated.  A third ECG was recorded another 20 minutes later:

Now there is clearly sinus tachycardia.  There is an incomplete RBBB.  There is unusual ST elevation in V1-V3 which does not look like STEMI.

No charts had yet been found.  At this point in time, the cardiologist was called and he recognized the patient and stated that he has HOCM.  He was in favor of assessing the coronary arteries, and so the cath lab was activated.

The patient became more hemodynamically stable.  Another bedside echo was done:

This is a parasternal short axis, and shows very thick concentric hypertrophy, but better LV filling now, with much more effective cardiac output

Here is the parasternal long axis view:

An arterial line was placed, and the BP by arterial line was 190/120, with a heart rate of 130.  O2 saturations fell and the chest x-ray revealed pulmonary edema.  Fluids were stopped and esmolol was rebolused and the infusion restarted.  The BP improved at 130/80 with a pulse of 90-120

As access for a cooling catheter was difficult, it was decided to delay targeted temperature management until after the angiogram.

Shortly before transfer to the cath lab, this ECG was recorded:
Sinus tachycardia with narrow complex, with delta waves.  ST elevation largely resolved.

It was learned that the patient had a history of HOCM and WPW, and also a history of severe embolic ischemic stroke due to paroxysmal atrial fibrillation, with hemorrhagic transformation.  Because of this bleeding danger, and also because the physician did not believe that the patient was having a acute coronary syndrome, no aspirin or Plavix or heparin was given.  The possibility was also considered that this was all initiated by a cerebral hemorrhage that caused stress cardiomyopathy.  He therefore underwent a CT scan of the head prior to angiography. This had no new findings, only the previous ischemic stroke (encephalomalacia).

The angiogram showed normal coronaries.  

Peak troponin I was 51 ng/mL (large Type 2 MI)

Formal echo showed HOCM with no outflow obstruction.

Cooling: The patient underwent targeted temperature management.

By 72 hours, the patient showed no signs of awakening, but by 96 hours was intermittently following commands.  By 7 days, the patient was "very interactive."

Learning Points:

1.   With excellent CPR technique, patients in ventricular fibrillation can be resuscitated even after a very long down time.  In this case, even with a left ventricle that could barely fill, the CPR was effective enough to have adequate perfusion.  Good chest compressions, at the right rate (at least 100) and depth (at least 5 cm, or 2 inches), decompression, ITD (ResQPod), slow ventilations (10/min), are among the many critical interventions that may lead to successful resuscitation.  Whether co-incidental or not in this case, we have had good rates of conversion of VF when esmolol is given.  

See this case of 68 minutes of cardiac arrest in a paramedic, plus recommendation from a 5 member expert panel on CPR.

2.  Not all cardiac arrest, even with pathologic ST elevation, is due to STEMI.   Cardiomyopathies, combined with cardiac arrest, can result in bizarre ECGs.  Stress cardiomyopathy may cause VF and ST elevation, and other PseudoSTEMI patterns may be present but unrelated to the VF.

3.   ECGs may be very bizarre immediately after defibrillation. Give a few minutes to record another before coming to conclusions.

4.  Bedside ultrasound is incredibly valuable in cardiac arrest, both for assessing cardiac function and for assessing carotid blood flow.

5.  Pulses may be absent when there is good perfusion through the carotid.  Use Doppler carotid ultrasound to assess carotid flow.  To my knowledge, there is no human literature on this.

6.  End Tidal CO2 is a good indicator of effectiveness of chest compressions. 
--By this systematic review in Resuscitation 2013, a value less than 10 mmHg (1.33 kPa) is associated with a very low return of spontaneous circulation.    
--In this systematic review from J Int Care Med 2014, the mean etCO2 in patients with return of spontaneous circulation (ROSC) was 26 mm Hg (3.5 kPa)

7.  Do not do any adverse neurologic prognostication prior to 72 hours after arrest, and it is preferable to wait even longer.  Here is one article from 2014 on this topic by Keith Lurie's group from HCMC and the U of Minnesota, and another (on which I am a co-author) from HCMC this summer of 2014.

8.  When a cardiac arrest victim has a history of "clots" and is on coumadin, one must entertain the diagnosis of pulmonary embolism.  However, ventricular fibrillation is an unusual presenting rhythm in pulmonary embolism:
--In this study, 5% of VF arrest was due to PE: V fib is initial rhythm in PE in 3 of 60 cases.  On the other hand, if the presenting rhythm is PEA, then pulmonary embolism is likely.  When there is VF in PE, it is not the initial rhythm, but occurs after prolonged PEA renders the myocardium ischemic.
--Another study by Courtney and Kline found that, of cases of arrest that had autopsy and found that a presenting rhythm of VF/VT had an odds ratio of 0.02 for massive pulmonary embolism as the etiology, vs 41.9 for PEA.