Monday, June 22, 2020

Diffuse T-wave inversions and a very long QT

This ECG was texted to me and I read it while mountain biking. 

My response was "takotsubo".

There is are diffuse T-wave inversions and a bizarre QT interval, very long
This is very typical of takotsubo.

Aside: I classify takotsubo ECGs into 2 broad categories:
1) ST Elevation which often mimics STEMI
2) T-wave inversion: does not mimic STEMI, but rather mimics NonSTEMI or reperfused STEMI

The next day I texted back to ask what the clinical presentation was, what the echo showed, and what the outcome was.

It turned out that this patient had severe alcoholism, alcohol withdrawal, and seemed to have had an alcohol withdrawal seizure prior to arrival.  His chief complaint was weakness.

Just days before, he had been in the hospital for alcohol withdrawal and had had this ECG:

ECG 2 (this is the first chronological ECG, 3 days before)
This shows both a very long QT and very prominent U-waves.

The patient at that time had suffered several episodes of torsades de pointes.  It turned out that he had a K of 2.5 and Mg of 1.3.  After these were replenished, another ECG was recorded:

ECG 3 (2nd chronological ECG, 2 days before)

Case continued:

Electrolytes were normal.

He had a bedside echo with normal function.

His peak troponin was barely elevated at 0.035 ng/mL (99th %-ile URL = 0.030)

So this was not takotsubo.  It was very unlikely to be acute MI.

Here is the final ECG:

ECG 4 (final chronologically)
T-wave inversions still present but not as pronounced.

The etiology of the T-wave inversions was not established.  Perhaps a transient occurrence after a seizure?  Or perhaps this was indeed Wellens' syndrome without a very elevated troponin.  Only an angiogram would show for certain.

MY Comment by KEN GRAUER, MD (6/22/2020):
I LOVE this case presented by Dr. Smith — as it touches on a number of important entities in emergency ECG interpretation. I focus My Comment on taking a closer look at these phenomena.
  • For clarity — I’ve preserved the numbering of tracings in this case in the sequence in which they are presented.
  • The case begins with ECG #1 — that was texted to Dr. Smith (Figure-1).

Figure-1: The initial ECG that was presented in today’s case (See text).

MY THOUGHTS on ECG #1: The rhythm is sinus at ~90/minute. The PR interval and QRS duration are normal. The frontal plane axis is slightly leftward (about -10 degrees). The rest of the tracing is replete with abnormalities!
  • As per Dr. Smith — the QT interval is markedly prolonged (it takes up more than 2/3 of the R-R interval).
  • There is voltage for LVH — satisfied by Peguero Criteria (ie, sum of deepest S in any chest lead + S in V4 ≥23 [female] or ≥28 [male]). In ECG #1 — the S wave in lead V3 (18) + the S in V4 (13) = 31.
  • NOTE: For review of the voltage criteria I favor for the ECG diagnosis of LVH — Please see the Table at the bottom of My Comment in the June 20, 2020 post in Dr. Smith’s ECG Blog.

But the most remarkable finding in ECG #1 is the presence of Giant T wave inversion. T wave inversion is seen in every lead except for leads aVR and V1.
  • Although some T wave inversion is common in many conditions — the term “giant T waves” is reserved for a select number of clinical entities that produce truly deep (ie, >5-10 mm amplitude) T wave inversion. Depth of the inverted T waves in leads V3 and V4 of ECG #1 exceeds 15 mm!
  • PEARL #1 — Truly “giant” T waves are not overly common. When you see them — it is worthwhile to consider the diagnostic entities listed in Figure-2. Without any history — it is impossible to know WHICH of the entities in this list is most likely. Once you know the history — Simply consider these diagnostic possibilities!
  • PEARL #2 — It is tempting on seeing the biphasic T wave with steep downsloping terminal component in lead V2 of ECG #1, to think of Wellens’ Syndrome. Wellens’ Syndrome is most probably not present in this case. While an acute ischemic coronary syndrome is a diagnostic possibility (ie, acute ischemia is included among the entities listed in Figure-2) — it is well to remember that there are other causes of the ST-T wave picture that we see in lead V2 of ECG #1. These include (among others) — LVH, cardiomyopathy, coronary reperfusion. In ECG #1 — I suspect the reason for the biphasic T wave in V2 is simply the reflection of the increased QRS and ST-T wave amplitude that we see, with lead V2 representing a “transition lead” placed between the tall positive T wave in lead V1 and the very deep negative T wave in lead V3.

Figure-2: List of diagnostic entities to consider when giant T wave inversion is seen (See text).

The 2nd half of this case covers an equally important series of Learning Points.
  • For clarity — I’ve placed the 2nd and 3rd tracings shown in this case together in Figure-3.
  • ECG #2 (that is shown in Figure-3) — was obtained during this patient’s previous recent hospitalization, which was for alcohol withdrawal. ECG #2 was obtained 3 days before ECG #1 — and at the time ECG #2 was obtained, the patient had severe electrolyte disturbance (serum K+ = 2.5 mEq/L and serum Mg++ = 1.3 mg/dl).
  • Among the other complications of this patient’s alcohol withdrawal during his recent hospitalization — were episodes of Torsades de Pointes and an alcohol withdrawal seizure.
  • ECG #3 (in Figure-3) — was obtained 1 day after ECG #2, at which time serum electrolytes had been corrected.

Figure-3: The 2 ECGs shown in this case from this patient’s recent hospitalization (See text).

MY THOUGHTS on ECG #2: As per Dr. Smith — ECG #2 manifests a number of findings characteristic of significant hypokalemia and hypomagnesemia. These include:
  • Marked sinus bradycardia and arrhythmia + a PVC (beat #8).
  • ST segment flattening in multiple leads, with subtle hint of ST depression in some of these leads.
  • Very prominent U waves in multiple leads (colored ARROWS in ECG #2).
  • A very long QT interval, that is seemingly further prolonged by prominent U waves that blend in with the terminal portion of the T wave in several leads.
  • PEARL #3 — We are told that this patient had several episodes of Torsades that occurred around the time that ECG #2 was obtained. This should not be surprising — since predisposing factors for developing Torsades include: i) Hypokalemia and hypomagnesemia; ii) Marked QT prolongation; and, iii) Marked bradycardia. Each of these factors were probably operative in this patient.
  • DID YOU NOTE? — that the U wave that occurs just before the PVC (beat #8) is clearly larger (PURPLE arrows) than all other U waves (BLUE arrows). I have no idea why this is so — but thought it interesting to point out.

MY THOUGHTS on ECG #3: As per Dr. Smith — ECG #3 was obtained after correction of the electrolyte disturbance that was present when ECG #2 was obtained a day earlier. Be SURE to note on comparing ECG #2 with ECG #3 that:
  • The rate of the rhythm in ECG #3 is now more appropriate — the rhythm is now regular — and there are now no PVCs.
  • The QT interval now looks normal.
  • There is now no more than minimal suggestion of tiny U waves (perhaps in leads V3, V4) — but there has clearly been marked reduction in U wave amplitude compared to ECG #2.
  • ST segment flattening remains in a number of leads — but the hint of ST depression we previously saw in a few leads has resolved.
  • PEARL #4 — The BEST way to get good at recognizing ECG signs of low K+ (and low Mg++) — is to profit from every opportunity you have when you know precisely what the low serum values of these 2 cations were on the initial ECG (in this case, in ECG #2) — WHEN you compare that initial tracing (when serum electrolytes were abnormal) to the ECG obtained after correction of the electrolyte disturbance (as we do in Figure-3).

The original question posed by assessment of ECG #1 in this case, was WHAT is the likely cause of the “giant” T waves that we see?
  • Figure-2 suggests the diagnostic entities to consider.
  • Of note — ECGs #2 and #3 (which were done 3 and 2 days before ECG #1) — did not show the picture of “giant” T waves that we see in ECG #1. Therefore — the likely cause would seem to be something that happened during the 2-day interim between when ECG #3 and ECG #1 was done.
  • Bedside Echo was normal, and peak troponin was barely elevated — which rules out Takotsubo cardiomyopathy, apical cardiomyopathy and acute ischemia from the list of potential causes in Figure-3.
  • I see nothing in the history suggestive of massive acute pulmonary embolism or Stokes-Adams attacks.

By the process of elimination — this leaves us with post-tachycardia syndrome and/or a severe CNS disorder — as the most likely cause(s) of these giant T waves.
  • This patient had a number of documented episodes of Torsades during his previous hospitalization. We are told that electrolytes were corrected at the time of his discharge — but I don’t see indication as to what repeat serum electrolytes were at the time ECG #1 was obtained. It would therefore seem possible that serum K+ and Mg++ may have dropped during the 2 days since ECG #3 was done — leading to additional episodes of Torsades that might have been prolonged enough to produce a post-tachycardia “memory” effect after spontaneous resolution.
  • We are told this patient had at least one alcohol withdrawal seizure. Severe CNS disorders are notorious for producing markedly abnormal (sometimes bizarre) ST-T wave changes.
  • A remarkable feature of the “giant” T waves seen in ECG #1 — is how markedly prolonged the QTc is. Both post-tachycardia syndrome and CNS catastrophes are known for producing this degree of marked QTc prolongation — so either (and/or both) of these disorders may be responsible for the ECG picture we see in ECG #1.


  1. Greetings from Guatemala. I'm a first year cardiology fellow. Congrats and thanks for how you develop and explain this case.

    1. Nice to hear from Guatemala!! Thanks for the comment! Some time if you want to write a bit about Cardiology in Guatemala, I would love to post something like that.

  2. Great Sharing Dr Steve Smith. Thank you, love to learn more

  3. Thank you so much for this explanation

  4. Thanks for the spectacular explanation. In my opinion, the possibility of cardiac memory T wave causing this global T wave inversion is somehow low. The reason is that cardiac memory effect turns the T wave axis to the axis of QRS of prior widening QRS rhythm (No matter intermittent VT, LBBB, pre-excitation or pacemaker). Therefore, in this case, if it's cardiac memory effect, meaning that the QRS of the prior widening QRS rhythm should be all negative except for aVR and V1 (since all leads except for aVR and V1 presents T wave inversion in this case). I wonder if the prior widening QRS rhythm is arised from ventricle, where at the ventricle could provide negative QRS in all leads except aVR and V1. Since Torsade de Pointes changes its vector from time to time. I don't know what axis will present in the T wave if Torsade causing cardiac memory effect. Need your expertise for explanation.

    1. THANKS for your comment! I’m afraid I don’t have any explanation for the point you raise. I would guess that since the axis continually shifts with Torsades — that it would be difficult to predict the morphology of a post-arrhythmia “memory effect” — but I don’t have any data on this.

    2. Thanks for replying. I never met case with cardiac memory effect caused by Torsade, either.

  5. Recently, I met an interesting case presented diffuse deep TWI ECG pattern (global deep TWI except for V1 and aVR) just like the case you sharing. The patient had episodes of status epilepsy. Therefore, propofol infusion was given. Several days after propofol infusion, ECG showed diffuse STE of shark fin sign in the anterolateroinferior leads. After reviewing several articles that refers propofol infusion syndrome may cause the lambda-like STE ECG pattern, which indicates a poor prognosis. We decided to discontinued propofol. The day after propofol discontinued, diffuse shark fin STE pattern changed into diffuse deep TWI just like the case you sharing. The depth of TWI then gradually subsided within days. Inspired by this case, I wonder if there were also propofol infusion then discontinued in the case you sharing?


DEAR READER: We welcome your Comments! Unfortunately — due to a recent marked increase in SPAM — we have had to restrict commenting to Users with a GOOGLE Account. If you do not yet have a Google account — it should not take long to register. Comments give US feedback on how well Dr. Smith’s ECG Blog is addressing your needs — and they help to clarify concepts of interest to all readers. THANK YOU for your continued support!

Recommended Resources