Wednesday, March 25, 2020

A patient with abdominal pain associated with alcohol withdrawal and alcoholic ketoacidosis

While at work, one my partners showed me this ECG of a 50-something woman with abdominal pain associated with alcohol withdrawal and alcoholic ketoacidosis.  There was no reported chest pain or SOB.

What do you think?

I said it "looks like takotsubo.  Electrolytes might contribute.  Are they back yet? (they were not).  I do not think this is a coronary event."

He asked why.

I responded: "bizarre T-waves, with T-wave inversion and extremely long QT.  The computer measures the QT at 506 ms, but it really is more like 560-580 ms, with a QTc of 600-620 ms.  This is not at all typical of ACS but very typical of takotsubo or stress cardiomyopathy."

So he did a bedside echo:

Here is the apical 4-chamber

You can clear see that only the base contracts, that the apex is bulging like an octopus trap

Here is a parasternal short axis view. 
The anterior wall is closest to the probe, and the lateral wall is on the right.

Decreased systolic function and anterior and lateral wall motion abnormality

A previous formal echo from 1 year prior had an EF of 68% and no wall motion abnormality.

The "apical ballooning" supports takotsubo.  However, the presence of a wall motion abnormality which corresponds to one coronary distribution (proximal LAD affecting anterior and lateral walls) is not typical, and puts a coronary etiology a bit higher on the differential diagnosis.

If it is coronary, then the artery is open (T-waves are inverted) and cath lab activation not necessary.

K returned at 3.2 mEq/L, Mg at 1.4 mg/dL (normal 1.6-2.6), Lactate at 7.7 mmoles/L and Troponin I at 0.219 ng/mL (URL = 0.030 ng/mL)

Another ECG was recorded 45 minutes later:
Slightly deeper T-wave inversion

Here is the overnight troponin profile:
This does not differentiate a Non Occlusion MI from Takotsubo

A formal echo was done in the morning:
The estimated left ventricular ejection fraction is 25%.
Decreased left ventricular systolic performance severe .
Regional wall motion abnormality-distal septum anterior and apex, anterolateral, lateral.

Stress induced cardiomyopathy (takotsubo-like LV dysfunction) probable.

Definite findings for very large regional wall motion abnormality, with segmental LV involvement c/w takotsubo type stress cardiomyopathy.  HOWEVER, obstructive CAD is not excluded as a contributing cause to the mechanism of regional LV dysfunction.

And a followup ECG:
Again, bizarre T-waves with very long QT
The computer measures the QT as 498 ms, QTc 534.
I measure it as at least 600 ms, possibly longer, with QTc of 635 ms

How can we differentiate ACS from takotsubo?

There are 2 general types of ECG presentation in takotsubo:

1. ST Elevation, often diffuse and not in one coronary distribution.

With this presentation, it is often necessary to obtain emergent angiography because what you think is takotsubo is often really STEMI/OMI.  Patients with occlusion of a wraparound (Type III) LAD that wraps around the apex such that it also supplies the inferior wall have diffuse ST Elevation and also have apical ballooning, such that you might think it is takotsubo.

See this case: 

Diffuse ST Elevation with Apical Ballooning: is it Takotsubo Stress Cardiomyopathy?

Examples of takotsubo with ST Elevation: 

Chest pain, sinus tachycardia, and ST Elevation

COPD exacerbation, what do the ECG and bedside echo show?

2. Bizarre T-wave inversions, in which case it could be ACS but does not need emergent angiogram to prove it.

See these cases of takotsubo with T-wave inversion:

Bizarre T-wave inversions, with Negative U-waves and Very long QT. And a myocardial viability study.


The cardiologists were sufficiently convinced of takotsubo that they did not do an angiogram.

Can one reliably differentiate the ST Elevation of Takotsubo from the ST Elevation of LAD occlusion?

The quote below is from a paper we wrote in Can J Cardiol a couple years ago:

David F. Miranda, M.D, Angie Lobo, M.D, Brooks Walsh, M.D, Yader Sandoval, M.D,
Stephen W. Smith, M.D.  
New Insights into the Use of the 12-lead Electrocardiogram for Diagnosing Acute Myocardial Infarction in the Emergency Department.   
Canadian Journal of Cardiology 34(2):132; February 2018.  Issue: Advances and Controversies in Cardiac Emergency Care. 

Takotsubo Cardiomyopathy

Takotsubo may manifest T-wave inversion, but also STE that mimics STEMI.  Although earlier work had suggested that ECG criteria might distinguish this STE from anterior STEMI,(57) recent literature does not support this result.(58 59) Although the specificity of various combinations of ECG elements for Takotsubo may be greater than 95%, the positive predictive value may be as low as 67% due to the low prevalence of Takotsubo.  Many anterior STEMI, especially due to wraparound LAD to the inferior wall, have similar ECG findings and also apical ballooning.(60)   Therefore, coronary angiography is often essential to rule out acute coronary occlusion, even when the STE pattern and cardiac ultrasound both suggest Takotsubo.

57. Kosuge M, Ebina T, Hibi K, et al. Simple and accurate electrocardiographic criteria to  differentiate takotsubo cardiomyopathy from anterior acute myocardial infarction. Journal of the American College of Cardiology 2010;55:2514-6.

58. Frangieh AH, Obeid S, Ghadri JR, et al. ECG Criteria to Differentiate Between Takotsubo (Stress) Cardiomyopathy and Myocardial Infarction. Journal of the American Heart Association 2016;5.

59. Vervaat FE, Christensen TE, Smeijers L, et al. Is it possible to differentiate between Takotsubo cardiomyopathy and acute anterior STEMI? J Electrocardiol 2015;48:512-9.

60. Mugnai G, Pasqualin G, Benfari G, et al. Acute electrocardiographic differences between Takotsubo cardiomyopathy and anterior ST elevation myocardial infarction. J Electrocardiol 2015;48:79-85.

Annotated Bibliography (relevant quotes from abstracts)

STEMI vs. Takotsubo:

57. Kosuge et al.:
"The absence of abnormal Q waves, absence of reciprocal changes, presence of ST-segment elevation in lead −aVR (i.e., ST-segment depression in lead aVR), and absence of ST-segment elevation in lead V1identified TC with sensitivities of 42%, 94%, 97%, and 94%, specificities of 74%, 49%, 75%, and 71%, and predictive accuracies of 71%, 53%, 77%, and 73%, respectively." 

58. Frangieh et al.:
"When comparing STEMI and STE-TTC, ST-elevation in –aVR (ST depression in aVR) was characteristic of STE-TTC with a sensitivity/ specificity of 43% and 95%, positive predictive value (PPV) 91%, and a negative predictive value (NPV) 62%. 0 .001= ST depression in aVR is accompanied by ST-elevation in inferior leads, sensitivity/specificity were 14% and 98% (PPV was 89% and NPV 52%) (P=0.001), and 12% and 100% when associated with ST-elevation in anteroseptal leads (PPV 100%, NPV 52%) (P less than 0 .001) On the other hand, STEMI was characterized by ST-elevation in aVR (sensitivity/specificity of 31% and 95% P less than 0 .001.

59. Vervaat et al. assessed these previous criteria:
"The existing ECG criterion was less accurate (76%) than in the original study (95%), with a large difference in sensitivity (26% vs. 91%). Only a frontal plane ST-vector of 60° could significantly distinguish TC from all acute anterior STEMI subgroups (p < 0.01) with an overall diagnostic accuracy of 81%. The mean amplitude in inferior leads II and aVF was significantly higher for patients with TC compared to all patients with acute anterior STEMI (p < 0.01 and p < 0.05 respectively) and the mean amplitude in the precordial leads V1 and V2 was significantly lower compared to proximal and mid LAD occlusion (p < 0.01)."

60. Mugnai et al.
"The absence of abnormal Q waves, the ST depression in aVR and the lack of ST elevation in V1 were significantly associated with TC (respectively: 52% vs 18%, p = 0.01; 47% vs 11%, p = 0.01; 80% vs 41%, p = 0.01). The combination of these ECG findings identified TC with a specificity of 95% and a positive predictive value of 85.7%."

ACS with T-wave inversion vs. takotsubo

Kosuge et al.  (Differences in Negative T Waves Between Takotsubo Cardiomyopathy and Reperfused Anterior Acute Myocardial Infarction)  compared the negative T-waves of Takotsubo to those of anterior MI and found that "negative T waves in lead -aVR (ie, positive T waves in lead aVR) and absence of negative T waves in lead V1 identified takotsubo with 94% sensitivity and 95% specificity."  In other words, takotsubo in this study had positive T-waves in both aVR and V1.

MY Comment by KEN GRAUER, MD (3/25/2020):
Interesting case that serves to review an important concept in Emergency Care ECG interpretation.
  • For clarity — I show the initial ECG in this case in Figure-1.
  • ECG #1 was obtained from a 50-something woman with abdominal pain associated with alcohol withdrawal and alcoholic ketoacidosis. No chest pain or dyspnea.

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

Dr. Smith has concisely summarized the KEY points ECG #1:
  • “Bizarre T waves, with T wave inversion and extremely long QT. The computer measures the QT at 506 msec — but it really is more like 560-580 msec, with a QTc of 600-620 msec."
  • Dr. Smith continued, “This is not at all typical of ACS — but it is very typical of Takotsubo or Stress Cardiomyopathy."
  • Bedside Echo then confirmed this diagnosis (Echo shown above by Dr. Smith).

MY THOUGHTS: I agree completely with the above interpretation of ECG #1 by Dr. Smith. For clarity — I think it worthwhile spending a moment to note the specific findings in ECG #1:
  • The rhythm is most probably sinus, with a heart rate of ~80/minute (albeit P wave amplitude is reduced, with an unusual P wave shape that could be from another site in the atria). Regarding intervals — the PR interval is probably normal (albeit it measures barely 0.12 second in lead II). The QRS complex is narrow (ie, not more than 0.10 second). As per Dr. Smith — the QT interval is markedly prolonged! I measure the QT to be ~560 msec in lead V3 — which given the heart rate ~80/minute, corresponds to a QTc ~620 msec.
  • The mean frontal plane QRS axis is somewhat vertical, at about +80 degrees.
  • Voltage criteria for LVH are met in several leads. These include: i) Deepest S wave in V1,V2 + tallest R wave in V5,V6 ≥35 mm; ii) R wave ≥18 mm in lead V6; andiii) R wave ≥20 mm in any inferior lead (Please see Extra Figure-A below for the user-friendly ECG criteria I favor for diagnosis of LVH).
  • Deep Q waves are seen in multiple leads (ie, leads II, III, aVF; V4, V5, V6 — with a smaller Q wave in lead I).
  • ST segment coving with slight elevation — that leads into deep, symmetric T wave inversion, is seen in high lateral leads I and aVL — and in the anterior leads (especially in leads V2 and V3).

IMPRESSION of ECG #1: As per Dr. Smith — the etiology of ECG findings in this tracing may be multifactorial:
  • Hypokalemia and/or hypomagnesemia are common causes of QT (QU) prolongation — and both of these electrolyte levels came back low (serum K+ = 3.2 mEq/L; serum Mg++ = 1.4 mg/dL in this patient who was acidotic). That said, the fact that the QTc remained extremely prolonged on the follow-up ECG (by which time serum electrolyte status and the patient’s acidosis should certainly have improved) — tells us that these metabolic abnormalities probably contributed little to this initial ECG picture.
  • The ample QRS voltage in this patient with a long history of alcohol abuse increase the likelihood that she had a baseline alcoholic cardiomyopathy — and that disorder could clearly contribute to some of the ST-T wave abnormalities seen in ECG #1 (Unfortunately, no baseline ECG was available for comparison). That said — the markedly prolonged QTC, with ST coving and deep T wave inversion in anterolateral leads is not what we’d expect from alcoholic cardiomyopathy alone.
  • Instead (as per Dr. Smith) — the clinical history of this acutely ill patient with the initial tracing shown in Figure-1 is most suggestive of Takotsubo Cardiomyopathy.

CONFESSION: I find it easy to forget about Takotsubo Cardiomyopathy. I’ve noticed that other even experienced providers also have a tendency to overlook this disorder. In the hope of serving as a helpful reminder — I’ve summarized in Figure-2 the ECG Findings that might be expected in Takotsubo Cardiomyopathy (adapted from Namgung in Clin Med Insights Cardiol). The initial ECG in this case ( = ECG #1) shows many of these features:
  • Although there is slight ST elevation in a number of leads in ECG #1 — there are much most striking findings in this tracing. These include: i) ST segment coving with deep, symmetric T wave inversion in leads I, aVL — and in leads V1, V2, V3; ii) Marked QTc prolongation is present (ie, a QTc ~620 msec)andiii) Deep Q waves are seen in multiple leads (which could be longstanding given increased QRS amplitude in these same inferolateral leads — but about which we can’t be sure without availability of a prior ECG).

Figure-2: ECG Findings in Takotsubo Cardiomyopathy — adapted from Namgung in Clin Med Insights Cardiol (See text).

Finally, for the purpose of brief review — I’ve added in Figure-3 a few pages that summarize Clinical Features of Takotsubo Cardiomyopathy. Note:
  • The patient in this case was not an elderly woman — but the combination of acute illness with abdominal pain, alcoholic ketoacidosis, and alcohol withdrawal is clearly a situation predisposing to severe ongoing catecholamine elevation.
  • Troponins were elevated — albeit not nearly as much as one might expect from the extent of ECG abnormalities.
  • Clinical follow-up on how this patient did (including cardiac cath results) were not available to me at the time I wrote my comments.
  • BOTTOM LINE: We need to remember to consider Takotsubo Cardiomyopathy in our differential diagnosis in cases like this. Our THANKS to Dr. Smith for this reminder!

Figure-3: Brief review of clinical features of Takotsubo Cardiomyopathy (Excerpted from Grauer K — ACLS-PB, 2013).

Figure-3 (Continued): Clinical features (Excerpted from Grauer K — ACLS-PB, 2013).

Extra Figure-A: The ECG criteria that I favor for diagnosing LVH (For more on “My Take” regarding the ECG diagnosis of LVH — CLICK HERE).


  1. Steve and Ken...

    Very interesting and instructive case with excellent reviews by both of you. And thanks for the references! I hope you and your families are well.

    1. THANKS Jerry. We are fine in Gainesville, Florida — essentially at home almost all the time. That said — I was born and raised in NYC (across the street from Elmhurst Hospital that had a nightmare day yesterday with multiple ICU deaths ...). This WILL pass — but unfortunately still a very LONG way to go ... BE WELL!


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