Monday, October 26, 2020

Fascinating case of dynamic shark fin morphology - what is going on?

 Case submitted by Magnus Nossen MD from Norway, written by Pendell Meyers

A man in his 50s with no pertinent medical history suffered a witnessed cardiac arrest. EMS found the patient in VFib and performed ACLS for 26 minutes then obtained ROSC. 12 minutes later, the patient went back into VFib arrest and underwent another 15 minutes of resuscitation followed by successful defibrillation and sustained ROSC. In total, he received approximately 40 minutes of CPR and 7 defibrillation attempts. 

Here is his first ECG recorded after stable ROSC:

Originally recorded in 50 mm/s (the standard in Norway), here converted to 25 mm/s. There is sinus rhythm with a relatively normal QRS complex followed by a hint of STD that is maximal in V4-V6, which appears to me to be an expected amount of supply-demand mismatch ischemia from his immediately post-ROSC state. I do not see clear evidence of OMI or reperfusion at this time.

The patient was transferred immediately for angiogram which revealed no significant CAD, and no intervention was performed. The patient received therapeutic hypothermia at 33 degrees C for 24 hours. Echo revealed morphologically normal left and right ventricles with normal ejection fraction and no wall motion abnormalities. Troponin T peaked at 330 ng/L, then trended down. After rewarming he was noted to be neurologically intact. An ICD was placed due to suspicion of a primary arrhythmia. MRI showed a small area of the lateral wall that could potentially represent infarct. 

Several weeks later the patient presented with epigastric pain and lightheadedness. ICD interrogation showed episodes of polymorphic VT. After cessation of VT, bizarre QRS morphology was noted and is shown below. Repeat echo showed no wall motion abnormalities. He was started on amiodarone and admitted for telemetry.

While on telemetry the patient remained asymptomatic, however recurrent episodes of ECG abnormalities were recorded. These episodes are characterized by rapidly dynamic ST segment and T wave abnormalities that rapidly change from normal to hyperacute T waves to shark fin morphology and back to normal. This progression occurs and resolves rapidly, within seconds on telemetry, with clear beat-to-beat changes. The telemetry alarm states "ventricular rhythm" but it is clearly massive ST segment elevation. All of these episodes occurred without any symptoms reported from the patient, even after pointed questioning during the telemetry events.

Here are the telemetry recordings (each line is a long rhythm strip, followed immediately by the line underneath it):

This event lasts two minutes and shows rapid evolution from normal to hyperacute T waves, to terminal QRS distortion, to STE, to a progressively altered terminal portion of the QRS with STE and QRS morphology worsening in tandem. By the halfway point of the two minutes, the rhythm strip shows complete shark fin morphology, which then resolves abruptly in the last rhythm strip over the course of seconds. This is literally the OMI progression caught in action, but with the twist of adding the shark fin morphology to the classic OMI progression.

This event similarly starts with a progression from normal to enlarging T waves with simultaneously increasing STE. In this event the QRS retains a more normal, narrow appearance until longer in the progression, allowing us to see obvious huge STE and terminal QRS distortion still with a narrow QRS. During the fourth line of the telemetry image, the shark fin appearance re-emerges. During the fifth line the rhythm changes likely to a brief run of polymorphic VT, then quickly back to sinus with rapid resolution of all findings. 

He had many 12-lead ECGs obtained during various episodes (keep in mind that these ECGs are recorded using 50 mm/s):

Dr. Nossen comments on his thought process at this point: "With this rapid ST-E development and regression there was in my mind only one plausible explanation. Coronary spasm causing massive current of injury with shark fin ECG. I suspect LAD or LM. I would not expect ST-E to vanish in four beats with dissolving thrombus (also we know that the coronaries were clean). Immediately restored blood flow as with cessation of spasm I speculated is the only explanation for the fast resolutionThe patient was not experiencing any typical coronary chest pain at these short episodes of ST-E. My understanding of this is that the transmural ischemia is sudden, massive and triggering a very pro-arrhythmic state before ischemic symptoms appear. Thus presenting clinically with malignant arrhythmia without the classic ischemic symptoms."

The patient was referred for a provocative test for coronary spasm based on these findings. While awaiting testing the patient was re-admitted due to short episodes of chest pressure. Telemetry again showed the same shark fin appearance multiple times. He was started on treatment for suspected coronary spasm. One incident was complicated by multifocal PVCs and a short episode of polymorphic VT, which self-terminated with regression of STE. 

The patient was sent for provocative angiogram which revealed significant spasm of the LAD, shown in the video below.

Here is the LAD before provocative testing:

Here is the mid LAD spasm induced during the procedure:

Here is a still image of the LAD spasm:

The patient is now doing well on long acting nitrates and diltiazem. ICD interrogations have been unremarkable. The patient underwent another ambulatory 48 hour holter ECG which showed only sporadic PVCs with no evidence of shark fin morphology or STE. He has been asymptomatic and doing well since.

Learning Points:

The myocardium doesn't know the etiology of OMI (ACS, spasm, dissection, embolus, etc.), and all acute coronary occlusion causes may cause the same ECG findings and progression of OMI. This reminds us that the ECG is a surrogate for what we are actually trying to figure out, and it cannot be expected to differentiate the etiologies of OMI. That said, ACS is by far the most common and treatable cause.

Serial ECGs and continuous 12-lead telemetry can be very helpful for both diagnosis and for learning the ECG.

When OMI is extremely brief, there may not be any clear ischemic symptoms during the brief episodes, and there may be no large troponin increase measured, and the ECG may not exhibit the classic reperfusion sequence after the episodes as they would for an OMI that was more sustained and cause significant MI.


MY Comment by KEN GRAUER, MD (10/26/2020):


Superb discussion by Dr. Meyers of this fascinating case captured in intricate detail by Dr. Magnus Nossen.

QUESTION: How many of YOU recognized the ECG format used for the 1st tracing shown above in this case? Please TAKE another LOOK (Figure-1):

  • This was the 1st ECG recorded following ROSC after a long course of ACLS. We are told that this ECG was originally recorded at 50 mm/sec speed — but had been converted to 25 mm/sec. But — WHAT about the limb lead arrangement?

Figure-1: The initial ECG shown above in today’s case (See text).

POINT #1: The ECG shown in Figure-1 has been recorded using the Cabrera Format. Electrocardiography is an international tool. While a standard ECG format (with no more than minor variation) is used throughout the United States — variations in format are used in some other countries.

  • A number of ECG parameters may vary in these other ECG formats. These most commonly include the speed of recording (ie, use of 50 mm/second — instead of the standard 25 mm/second speed used in the U.S. and — the sequence of the limb lead display (which is different in the Cabrera format — as shown in Figure-1).
  • Because we become accustomed to whatever ECG format is used in the country in which we practice — other ECG formats may be unfamiliar to us. As a result — the limb lead layout for the ECG shown in Figure-1 may have seemed strange when you first saw the initial ECG in today’s case. This is because the 1st limb lead displayed in Figure-1 is lead aVL  and, an aVR display occurs between leads I and II.

The Cabrera Format:

Initial description of the sequential limb lead format shown in Figure-1 was made by Fumagalli in 1949. Despite this — credit for this format is attributed to Cabrera. The Cabrera system has been in routine use in Sweden since 1977. A number of other countries also use the Cabrera format (ie, witness today’s case, contributed by Dr. Nossen from Norway). There are 2 Modifications in the Cabrera Format compared to the standard format used in the United States and in many (most) other countries: i) A recording speed of 50 mm/second is used in the Cabrera format; andii) A different limb lead sequence is used.

  • As I “confessed” in the September 26, 2018 post of Dr. Smith’s ECG Blog — my brain is “programmed” to interpreting 12-lead ECGs and rhythm strips at the 25 mm/second speed that is standard in the United States. After 4+ decades of interpreting tens of thousands of tracings — there is an instant (automatic) process of “pattern recognition” that I find takes place in my brain, even before I begin systematic interpretation of any given tracing. This process is invalidated by the unfamiliar appearance of different-sized complexes that are produced when a 50 mm/second recording speed is used.
  • POINT #2: Other countries (such as Germany) often use a 50 mm/second recording speed. Usually it will be obvious on sight when a 50 mm/second speed has been used — but sometimes it won’t be. This is especially true when assessing narrow QRS rhythms for heart rate (ie, a narrow QRS rhythm may appear widened and excessively slow if you fail to recognize a 50 mm/second recording speed). Therefore — it’s important to be aware of the recording speed (and to clarify if the tracing you are interpreting is from a foreign country — especially if the recording speed isn’t marked at the bottom of the ECG).

The Altered Limb Lead Sequence in the Cabrera Format:

The insert in the lower right-hand portion of Figure-2 shows the rationale for the altered limb lead sequence in the Cabrera format.

  • Rather than using lead aVR — the Cabrera format uses negative aVR ( = lead -aVR) — which is situated directly opposite (ie, 180 degrees away) from positive aVR. As shown in the insert of Figure-2  lead -aVR is situated at +30 degrees (within the BLUE rectangle).
  • In many ways — the Cabrera Format offers a much more logical display of limb lead sequencing. As opposed to the traditional U.S. format (in which limb leads are grouped into standard leads I,II,III — and augmented leads aVR,aVL,aVF) — there is gradually progressive (equally spaced) sequencing with the Cabrera format, beginning with the most superior lead viewpoint = lead aVL (at -30 degrees) — and moving gradually rightward (at 30 degree intervals) — until finally arriving at the most rightward placed lead = lead III (at +120 degrees).
  • The Cabrera format enhances the clinical utility of aVR — by effectively adding lead –aVR as a transition lead between lateral and inferior frontal plane location. This allows greater specificity for determining the extent of high lateral and inferior lead ischemia or infarction. It also simplifies both axis and ST-T wave vector calculation in the frontal plane — since no more than a glance at the 6 sequential Cabrera leads is now needed for instant determination of which lead(s) manifests greatest net QRS and/or ST-T wave deflection. For example, in Figure-2 — Note gradual transition of the ST segment straightening with slight depression that actually begins with lead -aVR instead of with lead II, as would have been suggested by the U.S. standard lead format.
  • Another potential advantage of the Cabrera format for limb lead sequencing — is that comparison with serial tracings in a given patient is easier. This is because the more gradually progressive Cabrera format for limb lead sequencing makes serial variation in Q wave presence, QRS amplitude, and ST-T wave displacements much more evident as to what represents probable “real change” — vs change in ECG waveforms that is more likely the result of some change in lead placement.
  • BOTTOM Line: Despite the above potential advantages that may be derived from use of the more logical limb lead sequencing of the Cabrera format — it seems unlikely to replace the non-sequential traditional U.S. format, at least for the immediate future. Old established habits are difficult to break ... — even when a newer approach seems technically easy to implement and clinically advantageous.

Figure-2: The insert that I've added to Figure-1 (in the lower right portion of this Figure) — shows the rationale for limb lead sequencing used in the Cabrera format (See text).

How I Interpret ECGs recorded at 50 mm/second Speed:

Ever increasing use of the internet has changed medicine, not to mention our life. Literally, hundreds of thousands of providers frequent the numerous ECG forums available on the internet — which means that whenever we participate in internet ECG case discussions — we are interacting with medical provider colleagues from all over the world. Tracings using both 25 mm/second and 50 mm/second recording speeds — as well as Cabrera and other formats will be posted. So it’s important to be aware that we may encounter a variety of ECG formats.

  • When I first received Dr. Nossen’s submission of today’s case — I noted that many of the multi-lead telemetry recordings and 12-lead tracings he sent were in Cabrera format with 50 mm/second recording speed (ECG #1 was one of the exceptions — which Dr. Nossen had already converted for us to 25 mm/second speed).

Because of the difficulty that I’m happy to acknowledge regarding the interpretation of arrhythmias and 12-lead ECGs recorded at 50 mm/second speed — I wanted to share a quick and easy way I’ve developed for “visually correcting” the wider appearance produced by a doubling of recording speed to 50 mm/second.

  • I’ve labeled the 6th ECG shown above in today’s case as ECG #6. This multi-lead monitor recording uses Cabrera limb lead sequencing and a 50 mm/second recording speed (LEFT panel in Figure-3). Especially because of the shark-tooth ST distortion during the first half of the recording — I initially had trouble conceptualizing what I was seeing.
  • PEARL: To compensate for the 50 mm/second recording speed — I uploaded this tracing to Power Point and: i) I reduced the width of this tracing by 50% (ie, from 10 inches to 5 inches); andii) I unchecked the “Keep Proportions” ( = “Lock aspect ratio” box) in the Format Pane under Size options in Power Point. Doing so narrows the width of the ECG by 50% without affecting the height. As can be seen in the RIGHT panel in Figure-3 — You’ll get twice as many little boxes (which must be considered when assessing rates and intervals) — but QRS complexes and ST-T waves now look "normal" to my eye.

Figure-3: I’ve reproduced the 6th tracing shown above in today’s case, first in Cabrera format recorded at 50 mm/second speed (LEFT panel) — and then in the RIGHT panel, after “compensating” for 50 mm/second speed by reducing width of the tracing by 50% without altering tracing height (See text for full explanation).


NOTE: My sincere THANKS to Dr. Magnus Berger Nossen (Norway) for sharing the tracings and this case with us!



  1. Post-Soviet countries also often (but not always) use 50 mm/s speed. But more annoying is "automatic" Y scale on some devices, switching between 10 and 5 mm/mV in different lead groups (usually 10 in limb leads and V1..V3, then suddenly 5 in V4..V6).

    1. @ Alexey — THANK YOU for your comment. I was not aware of the “automatic” Y scale that switches between 10 and 5 mm/mV in different lead groups in certain post-Soviet countries. However, I was aware that on many prehospital ECG recording systems in the U.S. — QRS amplitudes are automatically truncated. Potential adverse consequences that may result from this automatic amplitude truncation were discussed in My Comment at the bottom of the page from the February 6, 2020 post in Dr. Smith’s ECG Blog (GO TO — ).

    2. >>I was not aware of the “automatic” Y scale that switches between 10 and 5 mm/mV in different lead groups in certain post-Soviet countries.
      I've seen this on my friend's ECG several weeks ago. Leads I..III, aVR..aVF and V1..V3 were recorded at 10 mm/mV, and V4..V6 suddenly at 5 mm/mV (and everything at 50 mm/s!)...

  2. a most amazing case. frightening in a way. illustrates with painful clarity (via Dr. Nossen's imaging) how "clean arteries" can lead to OMI, and even sudden death from malignant arrhythmia. and also how a life threatening event and disease may be missed between episodes of spasm.
    thank you so very much, Pendell, Magnum, and Ken!
    ps: i wonder if another option might be to stent across the spastic segment of artery.

  3. What a great case!
    I agree with the brilliant Dr. Nossen that the rapid resolution of ECG changes after each episode is most consistent with coronary spasm.
    One question that I find intersting is: Why is there such a rapid onset of ECG changes compared to classic OMI caused by thrombus? After all the occlusion caused by the spasm here is sub-total.

    Airway, Again big thanks to Dr. Nossen, one of Norway's leading authorities on ECG interpretation, and National Blitz Chess Champion.

    1. I agree it is fascinatingly fast. Presumably the spasm mechanism can close and open very rapidly, much faster than the average thrombosis then autolysis of OMI I suppose? Ultimately I don't have great data for your question because this is so rare. As for whether OMI is subtotal or not, we do embrace the idea that OMI can be subtotal and still cause the OMI progression. 19% of proven true positive STEMI have TIMI 3 flow at time of cath. Many more have TIMI 1 or 2. many times we think that this just represents some reperfusion between ECG and cath, but we also believe that a patient can have less than full occlusion and still undergo full thickness infarction during that subtotal occlusion. that's why in our definition of OMI we state "acute coronary occlusion, or near occlusion, such that imminent myocardial infarction will occur without reperfusion..." Hope that helps!

    2. OMI changes to happen fast! We just don't see it because we don't have continuous ECGs and especially not during angiograms.

      Here is a relevant article showing that STE can occur within 30 seconds of balloon inflation:

      Kenigsberg DN, Khanal S, Kowalski M, Krishnan SC. Prolongation of the QTc interval is seen uniformly during early transmural ischemia. J Am Coll Cardiol [Internet] 2007;49(12):1299–305. Available from:

      Prolongation of the QTc Interval Is Seen Uniformly During Early Transmural Ischemia

      David N. Kenigsberg, Sanjaya Khanal, Marcin Kowalski, and Subramaniam C. Krishnan

      Methods: To understand electrocardiographic changes accompanying early ischemia, ECGs recorded at 20-s intervals during angioplasty were analyzed with 2 automated systems: MUSE (n-74) and Interval Editor (IE) (n-50).

      Results: Ischemia prolonged Bazett’s QTc interval uniformly (100%). With MUSE, QTc interval prolonged from 423 ± 25 to 455 ± 34 ms (p < 0.001). With IE, QTc interval prolonged from 424.0 ± 27 to 458 ± 33 ms (p < 0.001). Time to maximal QTc interval prolongation, change in T-wave polarity, ST-segment elevation, and ST-segment depression were 22, 24, 29 and 35 s, occurring in 100%, 7%, 15%, and 7% respectively.

      Conclusion: QTc interval prolongs uniformly in ischemia. Compared with other ECG indexes, it is also the earliest.


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